Marketing
A path through The Default Library
========================================
How people decide what to buy, and how organisations earn a place in their minds honestly.


Step 1. Microeconomics: Why Nothing Is Free and Every Price Is a Message
------------------------------------------------------------------------
Why: Demand, willingness to pay and competition: the economics under every market.
Everything costs what you gave up, choices happen one step at a time, and a price is the signal that makes strangers cooperate. Until it is wrong, and then you can see exactly how wrong.

A coffee shop raises its flat white from 3.50 to 3.80 and, somehow, still looks busy. A ride-hailing app doubles its fares in the rain and, ten minutes later, there are more cars on the road. A city starts charging drivers to enter its centre, and the jam thins out without a single new rule about who may drive. Nobody commanded any of this. A price moved, and thousands of people each did a small, selfish, sensible thing.
  An economics graduate watches those three scenes and sees the same few ideas working: everything costs something you gave up, decisions happen one step at a time, and a price is a signal that tells buyers to hold back and sellers to show up. That toolkit is microeconomics: how individuals and firms choose, and what happens when those choices meet in a market. It is also the friendliest door into the whole subject, because you already use it every time you shop.

  
    The course in one box
    
      Who takes itEveryone on an economics, business or marketing degree, usually in year 1 as "principles" and again in year 2 as "intermediate microeconomics", with calculus. MIT puts its principles course (14.01) at the gateway to Course 14-1; business schools such as Wharton teach the same toolkit as "managerial economics".
      What it is forPredicting how people and firms respond when prices, taxes, rules or competitors change, and judging whether the result is good for anyone besides the people making the trade.
      Comes afterCalculus I (34/01): "one more" is a derivative.
      Leads toMacroeconomics (34/43), Game Theory (34/45), Behavioral Economics (34/46), International Trade & Finance (34/47), Public Economics (34/49), Strategy & Competition (34/54), Marketing Management (34/57) and Pricing (34/61).
      What this piece gives youYou won't be able to estimate a demand curve from real data. You will see why a price cap makes queues, why a tax costs more than it raises, and why a market with five firms behaves differently from one with one. Easier first reads: Where Money Actually Comes From (04/02), Pricing Without Fear (04/03) and Capitalism: How Markets, Prices and Private Property Organise an Economy (31/01). This one puts numbers and sliders on the same ground. All the markets below are illustrative, not data.
    
  

  Scarcity and opportunity cost: the real price is what you give up
  Wants have no ceiling; time, land, labour and money do. That gap is scarcity, and it makes every choice a trade. The honest price of anything is its opportunity cost: the best thing you gave up to get it. A two-hour film costs the ticket plus whatever else those two hours could have been. A university degree's biggest cost is often the wages you didn't earn.
  Economists draw this as a production possibilities frontier: the most of one good you can make for each amount of another. It bows outward because resources aren't equally good at everything. The first fields moved from wool to wheat are the ones best suited to wheat; the last fields are poor wheat land. So each extra unit costs more than the one before. This rising cost, not any law of nature, is what lets prices carry information.
  
    Interactive · A farm's frontier
    Wheat grown
      
    
    
    Illustrative farm: wool = 100 − wheat²/100. The cost of the next unit is wool(now) − wool(after one more wheat).
  
  At 40 wheat the farm can have 84 wool. One more wheat costs 0.81 wool. At 80 wheat the same extra unit costs 1.61 wool, twice as much. Nothing got worse; you just ran out of the easy fields. Remember this shape whenever someone says "just do more of it".

  Thinking at the margin: only the next unit matters
  People rarely choose all or nothing. They choose one more: one more hour of study, one more loaf baked, one more floor on the building. The margin is the edge where the decision lives. The rule an economics graduate carries: keep doing something while its marginal benefit is at least its marginal cost, and stop when the next unit would cost more than it brings in.
  do more while  marginal benefit ≥ marginal costfor a seller facing a market price, marginal benefit is just the price
  A bakery sells loaves at a fixed price. Its cost of the next loaf rises as the oven gets crowded and staff go onto overtime (an illustration, not a real bakery). The slider below is the owner's decision. Watch the profit hump: the top is exactly where the cost of the next loaf meets the price.
  
    Interactive · How many loaves?
    Price per loaf
      
    Loaves baked
      
    
    
    Illustrative: total cost = q + 0.05q², so the next loaf costs about 1 + 0.1q.
  
  At a price of 6 the best output is 50 loaves and profit is 125. At 40 loaves the owner is leaving money on the table; at 60 she is baking loaves that cost more than they sell for. A higher price pushes the best output up: at 10 it is 90. Lining up each price with its best output is exactly how a supply curve is built.

  Supply and demand: the price that clears the market
  Demand says that, other things equal, buyers want less as the price rises. Supply says that sellers offer more as it rises (the bakery just showed why). Draw both against price and they cross at one point: the equilibrium price, where the quantity buyers want equals the quantity sellers offer. Above it there is a surplus and sellers cut prices; below it there is a shortage and buyers bid up. The price is pulled to the crossing without anyone setting it.
  The key habit: separate moving along a curve (the price changed) from shifting it (something else changed: tastes, income, input costs, a frost). Try three shocks in a market whose starting point is price 30, quantity 40.
  
    Interactive · Shift the curves
    
      Reset
      Demand rises (a fad)
      Supply falls (a frost)
      Both
    
    Demand shift (units at every price)
      
    Supply shift (units at every price)
      
    
    
    Illustrative: demand Q = 100 − 2P, supply Q = 2P − 20, then shifted sideways.
  
  A fad lifts both price (35) and quantity (50). A frost lifts the price to the same 35 but cuts the quantity to 30. Same price, opposite quantities: a graduate reads the quantity to tell whether demand or supply moved. When both shift by 20 the price jumps to 40 and the quantity barely changes, because the two effects on quantity cancel.

  Elasticity: how loudly buyers react
  A price rise always lowers the quantity demanded. The question that matters is how much. Price elasticity of demand is the percentage fall in quantity for a 1% rise in price. If it is below 1 in size, demand is inelastic: buyers shrug (salt, insulin, petrol in the short run). If it is above 1, demand is elastic: buyers walk (one brand of cereal, a restaurant meal).
  elasticity = % change in quantity ÷ % change in pricesellers' revenue rises with price when demand is inelastic, and falls when it is elastic
  A real case: from January 2014 Mexico taxed sugary drinks at one peso a litre, commonly described as roughly a tenth of the price. Purchases of the taxed drinks fell about 6% over 2014, and about 9% in the lowest-income households (Colchero and colleagues, BMJ, 2016). If the whole tax reached shelf prices, that is an elasticity around 0.6: buyers responded, but less than one-for-one. It is a rough reading, since pass-through to prices varied.
  
    Interactive · What does a 10% price rise do?
    
      Staple (0.3)
      Unit elastic (1)
      Restaurant meal (2)
      One brand (3)
    
    Elasticity (size)
      
    Price
      
    
    
    Illustrative constant-elasticity demand: Q = 100 × (P ÷ 10)−e. Real elasticities differ by market, time horizon and how many substitutes exist.
  
  With elasticity 0.5, a 10% price rise cuts quantity by about 4.7% and raises revenue by about 4.9%. At 2, the same rise cuts quantity by 17.4% and revenue by 9.1%. That is why a monopolist never charges where demand is inelastic: it could earn more by charging more.

  Surplus and the triangle nobody gets
  A trade happens only if it makes both sides better off. Consumer surplus is the gap between what a buyer would have paid and what they did pay. Producer surplus is the gap between what a seller received and the least they would have taken. In the market above, each is worth 400, so the total is 800. A competitive market, left alone and free of side effects, produces exactly the trades that make that total as large as it can be.
  Now interfere. A tax drives a wedge between what buyers pay and sellers keep, so some trades that were worth doing no longer happen. A price ceiling below the equilibrium makes sellers supply less than buyers want. Either way the lost trades are the deadweight loss: surplus that went to nobody. It is a triangle, so it grows with the square of the wedge.
  
    Interactive · A tax, or a price cap
    
      Tax per unit
      Price ceiling
    
    Tax per unit
      
    
    
    Same market as above (price 30, quantity 40, surplus 800). For the cap, the best case is assumed: the buyers who value the good most get it. Real rationing is usually worse.
  
  A tax of 10 shrinks the quantity from 40 to 30, raises 300, and destroys 50 of surplus. Double the tax to 20 and the deadweight loss does not double, it quadruples to 200, while revenue rises only to 400. A ceiling at 20 leaves 20 units on sale where buyers want 60: a shortage of 40, a loss of 200 even in the best case, and an invitation to queue, bribe or build a black market. Note what a tax can do that a cap cannot: it raises money that can be spent. The next piece in this thread, Public Economics (34/49), is about who ends up actually paying it.

  Market structures: from a thousand sellers to one
  "Supply and demand" assumes sellers are too small to move the price. That holds for a farmer selling wheat, not for an electricity grid or a search engine. Economists line up a spectrum. At one end is perfect competition: many sellers, identical goods, so each takes the price as given and price is driven to marginal cost. At the other is monopoly: one seller, who can choose the price, and who restricts quantity to keep it high. In between are oligopoly (a few firms who watch each other) and monopolistic competition (many firms with slightly different products, like restaurants).
  The simplest model that shows the whole spectrum is Cournot's, from 1838: firms choose quantities, each knowing the others do the same. The more firms, the more each fears flooding the market, and the closer price falls to cost.
  
    Interactive · How many rivals?
    
      Firms compete on quantity
      Perfect competition
    
    Number of firms
      
    
    
    Illustrative: demand P = 100 − Q, each firm's marginal cost 20. One firm sets 40 units at 60. Cournot with n firms: Q = 80n ÷ (n+1).
  
  One firm charges 60 and sells 40; two firms get 46.7; five get 33.3; twenty get 23.8, nearly the competitive 20. Buyers gain as firms are added (consumer surplus goes from 800 with one firm to about 2,900 with twenty), and the deadweight loss shrinks from 800 to about 7. This is the economic case for competition policy, and the reason Strategy & Competition (34/54) is a whole course. Entry matters more than the headcount, though: a market with one firm but easy entry can behave better than a market with three that are protected.

  Externalities and public goods: when prices miss something
  Prices work when the buyer and seller bear all the costs and benefits of the trade. An externality is a cost or benefit that lands on someone else: smoke from a factory, a neighbour's loud party, or, on the good side, a beekeeper's bees pollinating the orchard next door. The market quantity is too high for harms and too low for benefits, because the decision-maker ignores the third party.
  The classic fix is a Pigouvian tax (A. C. Pigou, 1920): charge the polluter the damage they cause, so the private price includes the social cost. London's congestion charge, which began on 17 February 2003 at five pounds a day (the standard rate is now £18 as of January 2026), is a price on road space for exactly this reason. Ronald Coase argued in 1960 that, where bargaining is cheap and rights are clear, the parties can sort it out themselves with no tax at all. Both are alive in every environmental debate.
  
    Interactive · A factory that pollutes
    Damage per unit made, to neighbours
      
    
      No policy
      Tax equal to the damage
    
    
    
    Same market as before. Making each unit costs the neighbours the amount on the slider.
  
  With damage of 10 per unit, the market still makes 40 units but the best quantity is 30. The ten extra units each cost society more than they are worth: a loss of 50. A tax of 10 moves the market to exactly 30. The other kind of miss is a public good: something nobody can be excluded from and that one person's use doesn't use up (a lighthouse, clean air, a vaccine's herd effect). Each person would like someone else to pay, so private markets undersupply it. When a resource is shared but rivalrous, a commons, Garrett Hardin's 1968 "tragedy of the commons" warned of overuse. Elinor Ostrom's field work, in Governing the Commons (1990) and her 2009 Nobel Prize, showed that communities from Swiss pastures to fisheries often build rules that avoid it without privatising or handing it to the state. The honest version is "it depends on the rules". More on this in Public Economics (34/49).

  What a graduate just sees
  
    Every choice has an opportunity cost, and it usually rises: the first unit is cheap, the thousandth is not.
    Decisions happen at the margin. Sunk costs and averages are the wrong numbers to look at.
    A price is information. When it moves, ask which curve shifted, and read the quantity to tell.
    Whether a price rise helps a seller depends on elasticity, which depends on substitutes and time.
    A price cap makes a shortage and a tax makes a wedge; the deadweight loss grows with the square of the wedge.
    "More competitors" and "freer entry" push price toward cost; fewer push it away.
    When someone else bears the cost or gets the benefit, the market quantity is wrong, and the fix is to make the price carry it.
    Markets can fail, but "the government fixes it" is a claim to check too: compare real alternatives, not an ideal against a market.
  

  Where it shows up next
  
    Macroeconomics (34/43): the same supply-and-demand logic, scaled up to the price level, interest rates and the whole labour market.
    Game Theory (34/45): when the firm count is small, strategy replaces the price-taking assumption. The oligopoly dial above is a simple game.
    Public Economics (34/49): externalities, public goods and who bears a tax become the whole subject.
    In the world: a surge price on a ride app, a carbon price, a minimum wage debate, antitrust cases against big platforms, and every "should we charge for this?" meeting in a company.
  

  If you want to go deeper
  
    📘 Textbook. N. Gregory Mankiw, Principles of Microeconomics (Cengage). The most widely assigned introduction in the US and much of the world: clear, short chapters.
    📘 Textbook. Daron Acemoglu, David Laibson and John List, Economics (Pearson), or Paul Krugman and Robin Wells, Microeconomics (Worth). Two modern alternatives: the first leans on evidence and experiments.
    📘 Gentler. Charles Wheelan, Naked Economics (W. W. Norton). Almost no maths; the ideas as stories.
    🌐 Free course. MIT OpenCourseWare 14.01 Principles of Microeconomics (Fall 2023): lectures and problem sets. Also CORE Econ's The Economy and Khan Academy Microeconomics, both free.
  

  
    The takeaway
    Every choice costs the best alternative you gave up, and it is made one step at a time. Prices carry that cost between strangers: when they are right, markets turn scattered selfishness into trades that make both sides better off. When they are wrong (a cap, a tax, a monopolist, a neighbour who bears the smoke), something is lost, and the skill is to find which wedge is the problem.
    Try this: this week, price one thing you buy in terms of what you gave up, and find your margin. Would you pay for one more? Then watch a price move at a shop or an app and ask which curve shifted.
  

  
    Sources & Further Reading
    Sources. The ideas are standard textbook microeconomics (Mankiw; Acemoglu, Laibson and List). Every number in the interactives uses invented, illustrative markets and was computed in node: farm frontier 84 to 83.19 (cost 0.81) at 40 wheat and 36 to 34.39 (1.61) at 80; bakery best output 50 with profit 125 at price 6, 90 at price 10; equilibrium 30 and 40, shocks 35 and 50 (fad), 35 and 30 (frost), 40 and 40 (both); elasticity 0.5 gives −4.65% quantity and +4.88% revenue, elasticity 2 gives −17.36% and −9.09%; tax of 10 gives quantity 30, revenue 300, deadweight loss 50, tax 20 gives 200 and revenue 400; ceiling 20 gives quantity 20, shortage 40, loss 200; Cournot with 1, 2, 5, 20 firms gives price 60, 46.7, 33.3, 23.8 and consumer surplus 800 to about 2,900; externality of 10 gives optimum 30 and loss 50. Mexico's 2014 sugary-drink tax: Colchero, Popkin, Rivera and Ng, BMJ 352:h6704 (2016), which reports a 6% fall in taxed-drink purchases in 2014 and 9% in the lowest-income households; the one-peso-a-litre rate is as reported in public summaries and the "about a tenth of the price" is approximate. London congestion charge: began 17 February 2003 at £5, £18 from 2 January 2026 (Transport for London, via Wikipedia). Coase 1960 (Nobel 1991), Pigou 1920, Hardin 1968 (Science), Ostrom 1990 (Nobel 2009), Cournot 1838 are standard history of economic thought.
    
      🌐 MIT OCW 14.01 Principles of Microeconomics
      🌐 CORE Econ, The Economy
      📘 Mankiw, Principles of Microeconomics
      📘 Wheelan, Naked Economics


Step 2. Marketing Management: Finding the Customer, Not Shouting at Them
------------------------------------------------------------------------
Why: Segment, target, position, and the marketing mix: the frame for everything else.
Most people think marketing is advertising. A marketing graduate sees something earlier and larger: who the customers are, what they would actually value, and where in their minds a product should stand.

A small bakery puts a sign on the pavement: "Fresh bread, 7 a.m." Nothing about that sign is clever. But somebody decided that people walking to the station want bread they can grab on the way, that they will pay a little more for it hot, that 7 a.m. is when they pass, and that the shop belongs on that street and not in a mall. The sign is the last, smallest decision in a chain of earlier ones.
  That chain is marketing management. Advertising is the part you can see. The rest, deciding which customers, what to offer them, at what price, where to sell it and how to say it, is done before a single poster is printed. Companies that skip the chain make beautiful ads for products nobody needed.

  
    The course in one box
    
      Who takes itEvery business student, usually in year 1 or 2 (programmes such as Wharton's and Kellogg's put it early), and every marketing major as the door to the rest. MBA programmes teach it in the first term. Engineers and designers often meet it as an elective.
      What it is forChoosing who you serve and what you promise them, then lining up product, price, place and message behind that choice.
      Comes afterMicroeconomics (34/42): demand, willingness to pay and competition.
      Leads toConsumer Behavior (34/58), Marketing Research & Analytics (34/59), Digital Marketing & Advertising (34/62), Product Innovation & Go-to-Market (34/63). Brand Management (34/60) and Pricing (34/61) follow Consumer Behavior. Entrepreneurship (34/56) needs it too.
      What this piece gives youYou will not be able to write a marketing plan for a bank. You will be able to look at any product and say who it is for, what it promises, and where it sits against its rivals. Easier first reads in the library: How to Talk to Customers (04/05), Pricing Without Fear (04/03) and the course Social Media Marketing. This one goes further and puts numbers on the ideas.
    
  

  What marketing is, and what it isn't
  The American Marketing Association's definition is a mouthful: marketing is "the activity, set of institutions, and processes for creating, communicating, delivering, and exchanging offerings that have value for customers, clients, partners, and society at large." Read it for the verbs. Creating and delivering come before communicating. Advertising is one verb out of four.
  The idea underneath is customer value: what the customer gets minus what the customer gives up. What they give up is not only the price. It is also time, effort, risk and the chance they chose wrong. A €2 coffee that takes fifteen minutes in a queue may deliver less value than a €3 coffee next door. Marketers say a customer buys when perceived benefits beat perceived costs, and better than the alternative does.
  Value = perceived benefits − perceived costscosts include money, time, effort and risk · the customer compares you with the next-best option
  Marketing also is not manipulation by definition, and not "selling". Peter Drucker (Management, 1973) put the ideal this way in 1973: to know and understand the customer so well that the product fits and sells itself. Whether firms live up to that is a separate question; Consumer Behavior (34/58) looks at what works on people and what doesn't.

  Segmentation: nobody buys for the average person
  Customers differ. Some will pay double for a feature; others only look at price. If you design for the average, you serve nobody well. Segmentation splits a market into groups whose members are alike in what they want and unlike other groups. Wendell Smith proposed the idea in the Journal of Marketing in 1956.
  You can cut by who people are (age, income, country), where they are, how they behave (heavy users, first-timers), or what they want (benefits sought). Textbooks add four tests for a useful segment: it can be measured, it is big enough to be worth serving, it can be reached, and you can act on it with a different offer. A segment of "people who feel optimistic on Tuesdays" fails them all.
  Here is segmentation done by a computer. 120 invented customers sit on two dials: how price-sensitive they are, and how much they want features. A standard algorithm, k-means, puts the points into k groups. Move k and watch what the grouping buys you.
  
    Interactive · Cut the market into k segments
    Number of segments (k)
      
    
    
    Illustrative data: 120 invented customers generated around four hidden groups (30 each, spread about 9 points). Horizontal: price sensitivity. Vertical: wanting features. The spread number is total squared distance from each point to its segment centre, as a share of the spread with one segment.
  
  With one segment, the "average customer" sits in the middle of the plot where almost nobody is. Going from 1 to 4 removes most of the spread; going from 4 to 6 buys little. That bend, the "elbow", is how analysts judge the number of segments worth having. The maths does not decide which segments the business can serve: that is a judgement, and a cost.

  Targeting and positioning: a place in the customer's mind
  Targeting is choosing which segments you will serve, and which you will let go. Positioning is the place you want in those customers' minds against the alternatives. The idea became famous through Al Ries and Jack Trout's 1981 book Positioning: The Battle for Your Mind. A position is relative: "cheaper than the leader", "the one for serious runners".
  Marketers draw it as a perceptual map: two attributes that matter to customers on the axes, brands as points, customers' ideal products as a cloud. A brand tends to win the customers closest to it. That gives a rule: a good position is near many ideal points and far from rivals. Drag your brand (or use the sliders) and see.
  
    Interactive · Place your brand on the map
    
    Your price position (left cheap, right premium)
      
    Your quality position (bottom basic, top high)
      
    Find the best open spotReset to the middle
    
    Illustrative: five invented rival brands (A to E) and 300 invented customer ideal points in three clusters. Each customer is assumed to pick the nearest brand (straight-line distance on the map). Real customers also care about habit, availability and trust, so treat the share as a thought experiment.
  
  The map shows a trade-off, not an answer. Where the customers are, the rivals are too, so you split the crowd. Where nobody is, you may win everyone nearby, but there may be nobody there. Press "Find the best open spot" and compare it with the middle: the search tells you where on this invented map a new brand would win most, and the number is only as good as the invented customers behind it.

  The marketing mix: four Ps, and the case against them
  Once you know whom you serve and the position, you have levers. E. Jerome McCarthy, in Basic Marketing (1960), grouped them as the four Ps: Product (what it is and does), Price (what it costs and how it is paid), Place (where and how the customer can get it) and Promotion (how you tell them). Neil Borden had popularised the phrase "marketing mix" in a 1964 article. The mix must fit together: a luxury watch at a discount-store price in a clearance bin sends two messages at once.
  The four Ps have lasted because they are easy to remember. They also have critics. Services (a haircut, a bank) add people, process and physical evidence, which Bernard Booms and Mary Jo Bitner turned into the seven Ps in 1981. Robert Lauterborn argued in 1990 that the Ps are the seller's view and should be flipped to the buyer's: customer solution, cost, convenience, communication (the four Cs). Others say the list is a checklist, not a theory, and ignores relationships, brands and digital platforms. Most textbooks, Kotler and Keller's included, still teach it as the starting frame and add to it.
  
    Seller's PBuyer's CThe question it asks
    ProductCustomer solutionWhat problem does it solve for them?
    PriceCostWhat does it really cost them, in time and risk too?
    PlaceConvenienceHow easy is it to get?
    PromotionCommunicationIs it a two-way conversation, not a broadcast?
  

  Customer lifetime value: what a customer is worth
  A bakery that thinks of a sale as €4 sees a €4 customer. A bakery that knows the customer comes back most weeks for years sees a customer worth hundreds. Customer lifetime value (CLV) is the total margin a customer is expected to bring across all the years they stay, in today's money.
  The simplest version needs three numbers: the margin a customer brings each year (m), the share who stay from one year to the next (the retention rate, r) and a discount rate (d) because next year's money is worth less than today's. If margin is collected at the start of each year the customer is active:
  CLV = m × (1 + d) ÷ (1 + d − r)expected years with you ≈ 1 ÷ (1 − r) · valid when r < 1 + d
  Worked: m = 120, r = 80%, d = 10% gives 120 × 1.1 ÷ 0.3 = 440. Raise retention to 90% and it jumps to 660. Five points of retention change the value far more than five points of margin, which is why firms obsess over keeping customers. Compare CLV with what it cost to win them (CAC): a common rule of thumb is that CLV should be a few times CAC, but it is a rule of thumb, not a law.
  
    Interactive · What is a customer worth?
    Margin per customer per year
      
    Retention rate
      
    Discount rate
      
    Cost to win the customer (CAC)
      
    
    Illustrative inputs. Model: margin paid at the start of each year the customer is active, constant retention and margin forever. Real customers churn unevenly, and CLV is a forecast, not a measurement.
  

  The funnel: a useful lie
  The funnel pictures customers falling out stage by stage: aware, interested, considering, buying, buying again. The idea is old: advertising writer E. St. Elmo Lewis is credited with an attention, interest, desire, action sequence around 1898, later called AIDA. A funnel is useful because stage rates multiply, so you can see where the big drops are and what a gain at one stage is worth at the end.
  
    Interactive · Walk 50,000 people through a funnel
    Visit (of people reached)
    Add to basket (of visitors)
    Buy (of basket)
    Buy again (of buyers)
    
    
    Illustrative: 50,000 people reached. Each stage rate applies to the one before.
  
  The funnel is a lie because people do not travel in a line. They loop back, skip stages, buy on impulse or never "consider" at all. Consultants at McKinsey drew a loop instead of a funnel in 2009, and the Ehrenberg-Bass school argues that most buying is far less deliberate than the funnel implies (Brand Management, 34/60, picks that up). Treat the funnel as a bookkeeping device, not a picture of a mind.

  Marketing myopia: you are in the customer's business
  The most reprinted article in the field is Theodore Levitt's "Marketing Myopia" (Harvard Business Review, July–August 1960). His claim: industries decline not because the market saturates but because managers define the business by the product, not by the need. His famous example was the railroads, which, he said, thought they were in the railroad business rather than the transportation business, and let aeroplanes and cars take the customer. A line Levitt popularised, "people don't want a quarter-inch drill, they want a quarter-inch hole", makes the same point in one sentence.
  The honest caveat: Levitt's story is a lesson, not a law. Railroads were heavily regulated, and a firm that stretches its definition too far ("we are in the energy business") can wander into things it does badly. The enduring part is the question to ask: what job is the customer hiring this for, and what else can do that job?
  Customers buy outcomes. Products are just the current way of delivering them.

  What a graduate just sees
  
    Every ad sits on top of a decision about who it is for. If you can't say whom, the ad is noise.
    "The average customer" is a fiction; there are groups, and serving one well means saying no to another.
    A product is a position: you can describe it by where it sits against rivals on what customers care about.
    Price, product, place and message have to tell one story. When they disagree, customers believe the price.
    Customers are worth years of margin, so keeping one is often cheaper than winning a new one.
    A funnel shows where the leaks are, but the customer's real path loops and skips.
    Ask what need a product serves, not what the product is; the need outlives the product.
  

  Where it shows up next
  
    Consumer Behavior (34/58): why people actually choose, which is rarely how the funnel says.
    Marketing Research & Analytics (34/59): how to find segments and measure what a campaign did.
    Brand Management (34/60) and Pricing (34/61): positioning made lasting, and the one P that brings money in.
    Digital Marketing (34/62) and Go-to-Market (34/63): the funnel and the mix on search, social and launch plans.
    In the world: product managers choosing a launch segment, a founder choosing a first customer, a nonprofit asking who a campaign is really for.
  

  If you want to go deeper
  
    📘 Textbook. Philip Kotler, Kevin Lane Keller and Alexander Chernev, Marketing Management (Pearson, 16th edition). The standard text for the course; best for its frameworks and cases.
    📘 Gentler. Seth Godin, This Is Marketing (2018): short, readable, about serving a small group well.
    🌐 Free course. MIT OpenCourseWare 15.810 Marketing Management (Sloan, Fall 2004, with lecture notes and cases).
  

  
    The takeaway
    Marketing starts long before the ad. Pick whom you serve, say where you stand against the alternatives, line up the product, price, place and message behind that, and measure a customer by years, not by the sale. Keep asking what need you are really serving.
    Try this: pick a product you use and write one sentence: "For [who], it is the [category] that [promise], unlike [rival]." If you can't fill the blanks, you've found a fuzzy position.
  
  
    Sources & Further Reading
    Sources. AMA definition of marketing (ama.org, panel of 2017). Smith, "Product Differentiation and Market Segmentation as Alternative Marketing Strategies", Journal of Marketing (1956). Ries & Trout, Positioning (1981). McCarthy, Basic Marketing (1960); Borden, "The Concept of the Marketing Mix" (1964); Booms & Bitner (1981); Lauterborn (1990). Levitt, "Marketing Myopia", HBR (1960; reprinted 2004). MIT OCW 15.810 page. Customer value, segmentation tests and CLV are standard textbook content (Kotler & Keller; Gupta & Lehmann on CLV). All data in the interactives is invented; the calculations (CLV 440 and 660; k-means spread; funnel counts) were computed in node. This is not business advice.
    
      🌐 MIT OCW 15.810 Marketing Management
      🌐 Levitt, Marketing Myopia (HBR)
      📘 Kotler, Keller & Chernev, Marketing Management
      📘 Godin, This Is Marketing


Step 3. Probability & Statistics: Reasoning with Uncertainty
------------------------------------------------------------
Why: Every survey, test and dashboard is statistics.
Nothing you measure is exact, and no part comes out the same twice. Statistics is the discipline of knowing how much to trust a number.

A machinist turns a hundred steel shafts to "10 millimetres". Measure them with a good micrometer and no two are the same: 9.981, 10.012, 10.004. Nothing is wrong. That scatter is a fact of every process, every material and every sensor.
  Now stack five of those parts end to end in a gearbox. Will the stack still fit the housing? A pollster calls a thousand people and says 52% will vote yes, give or take three. A drug trial says "significant". A forecast says 30% chance of rain. Every one of these is the same question: given scatter, how much should I trust this number?
  A graduate of probability and statistics doesn't see a single number any more. They see the shape of the scatter behind it. This piece builds that sight with eight things you can run: dice, tolerances, polls, a drug-trial trap, a belief that updates, and a gearbox stack.

  
    The course in one box
    Who takes it: almost every engineering and science student, usually in the second year; computer scientists meet it again before machine learning.
    What it's for: deciding what to believe, and how far to trust a measurement, when outcomes are uncertain.
    Comes after: Calculus I (34/01), which helps with continuous distributions, though the ideas here need little more than arithmetic.
    Leads to: Machine Learning (34/32) needs it directly. It also shows up in Materials Science & Manufacturing (34/13: process control and defects), Systems Engineering (34/22: reliability) and Computer Security (34/33: randomness and guessing).
    Honest line: you won't be able to design a clinical trial after this. You will see why "significant" doesn't mean "important", why a bigger sample beats a cleverer guess, and why most real quantities look like a bell curve. For easier first reads, see the library's Probability: How to Not Fool Yourself (12/03), Statistics for Reading the World (12/05) and Bayes' Theorem (12/17); this piece goes further into the engineering side.
  

  Random variables: a shape with a centre and a spread
  A random variable is a number that comes out of a process you can't predict one run at a time. The key move is to stop asking "what will it be?" and ask "what is the shape of what it could be?" Roll two dice and add them: 7 is six times as likely as 2 or 12, so the shape is a tent. Roll many times and the tent appears.

  
    Interactive · Roll dice, grow a shape
    
      Dice added together
      Roll 10 timesRoll 1,000 timesClear
      
      
    
    Without JavaScript: two dice add to 7 with probability 6/36 and to 2 or 12 with probability 1/36 each. The mean total is 7 and the standard deviation is about 2.42.
  
  Two numbers summarise a shape. The mean is its balance point: 7 for two dice. The standard deviation (the "spread") says how far values typically stray from the mean: about 2.4 for two dice, which means totals usually land between roughly 5 and 9. The variance is the standard deviation squared, and it is the one that adds up neatly when you combine independent things.
  standard deviation = √( average of (value − mean)² )
  Now the shape that matters most. Measured quantities (shaft diameters, heights, sensor noise) tend to form the normal curve, the bell. Its great convenience is a fixed rule: about 68.3% of values fall within one standard deviation of the mean, 95.4% within two, and 99.7% within three. Engineers use it every day. Say the shaft's spec is 10.00 mm ± 0.05 mm. Slide the process around and see what share of parts get thrown away.

  
    Interactive · How many parts fail the spec?
    
      Average diameter
      Spread (standard deviation)
      
      
    
    Without JavaScript: a centred process with standard deviation 0.025 mm (the spec limits are exactly ±2 standard deviations) scraps about 4.55% of parts. Cut the spread to 0.01 mm and about 0.00006% are scrapped. Shift the average 0.02 mm off-centre with spread 0.02 mm and about 6.7% are scrapped.
  
  Two lessons graduates keep. First, spread, not average, decides quality: a process centred perfectly but wide still scraps parts. Second, being off-centre costs a lot even when the spread is fine. That is why factories track both.

  The central limit theorem: why the bell keeps appearing
  Why should so many things be bell-shaped? The answer is one of the most useful theorems in mathematics. Take almost any random process, however lopsided, and average a few dozen independent draws. The averages pile up in a bell, centred on the true mean, with spread shrinking like 1/√n. Early forms are due to de Moivre in 1733 (for coin flips) and Laplace, whose general version dates from about 1810. Pick a lopsided source and drag the sample size.

  
    Interactive · Average anything, get a bell
    
      Flat (0 to 1)
      Very lopsided
      Coin (0 or 1)
    
    
      Draws averaged together: n
      
      
    
    Without JavaScript: for the lopsided source (mean 1, standard deviation 1) the average of n = 30 draws has standard deviation 1/√30 = 0.183, and the 3,000 averages form a close-to-normal bell around 1.
  
  At n = 1 you see the raw source. By n = 30 even the lopsided one is a near-bell. The prediction (the curve) uses only two numbers from the source, its mean and its standard deviation. The shape of the source is forgotten. That is why a measurement made of many small independent errors is normal, and why so much of statistics can pretend things are bell-shaped.
  standard error of an average = σ / √n
  In words: the average of n draws wobbles less than one draw, by the square root of n. To halve the wobble you need four times the data, not twice. It is the quietest expensive fact in all of experimental science.

  Sampling and confidence intervals: how sure is a poll?
  You can't ask everyone, so you ask n people and treat the share saying yes as an estimate. How far off might it be? The standard error from the last section gives the answer. For a share p the wobble is √(p(1−p)/n), and a 95% confidence interval is the estimate plus and minus about two of those. For n = 1,000 and a share near half, that is ±3.1 percentage points: the familiar "margin of error".
  The wording is subtle, and the interactive makes it concrete. "95% confidence" is a statement about the method, not about one interval: if you ran the procedure again and again, about 95 intervals in 100 would capture the true value. Here the truth is secretly 52%. Run 100 polls.

  
    Interactive · 100 polls, one truth
    
      People asked in each poll
      Run 100 polls
      
      
    
    Without JavaScript: with 1,000 people per poll the margin of error is about 3.1 points; run 100 polls and about 95 of the intervals cover the true value. With 4,000 people the margin halves to about 1.5 points.
  
  Watch two things. Roughly five in a hundred intervals miss the truth even when everything is done correctly. And quadrupling the sample from 1,000 to 4,000 only halves the interval. A poll of a country and a poll of a town need the same sample size for the same precision, as long as the sample is random. The real danger is not the margin of error; it is a sample that isn't random at all, which no formula fixes.

  P-values and Bayes: two ways to weigh evidence
  A hypothesis test asks whether data are surprising if nothing is going on. The p-value is the probability, assuming nothing is going on, of seeing a result at least as extreme as yours. Flip a coin 100 times and get 60 heads: for a fair coin the chance of a split at least that lopsided in either direction is about 5.7%. It is not the probability that the coin is fair, and it is not the probability that the result was "due to chance". The American Statistical Association issued a rare formal statement on this in 2016, with six principles. Two of them: p-values do not measure the probability that a hypothesis is true, and a p-value does not measure the size or importance of an effect.
  The classic trap is testing many things. At the usual cut-off of 0.05, a test of a true "nothing" comes out "significant" one time in 20 by luck. Test twenty unrelated things and the chance that at least one lights up is 1 − 0.9520, about 64%. Try it: each press tests twenty effects that are, by construction, nothing.

  
    Interactive · Test twenty things that do nothing
    
      Run 20 testsReset the tally
      
      
    
    Without JavaScript: twenty tests of pure noise at the 0.05 level give at least one false "discovery" about 64% of the time, and one in 20 tests on average.
  
  If you report only the one that lit up, you have a headline and no finding. This is part of why many published results fail to repeat: in the Open Science Collaboration's 2015 attempt to replicate 100 psychology studies, 97% of the originals had statistically significant results but only 36% of the replications did.
  Bayes' rule answers a different question: given the data, how should my belief move? Start with a prior belief, multiply by how well each possibility explains the data, and renormalise. Say you test parts and want the defect rate. Know nothing at first (any rate from 0% to 100% equally likely) and add parts one at a time.

  
    Interactive · Learn a defect rate from parts
    
      Add a good partAdd 10 good partsAdd a defective partReset
      
      
    
    Without JavaScript: after 20 good parts and no defects, a flat starting belief gives a most-plausible defect rate near 0%, an average of 4.5%, and a 90% credible range of roughly a quarter of a percent to 13%. Twenty clean parts cannot rule out a defect rate near one in eight.
  
  Frequentists and Bayesians disagree about the philosophy: one treats probability as long-run frequency and never assigns it to a hypothesis, the other as degrees of belief that data update. In practice both are used, often on the same project, and on big data sets they usually agree. The Bayesian version has one practical advantage here: after twenty clean parts it honestly says "we still can't rule out one in eight".

  Correlation, regression and the trap of causation
  Correlation (called r) measures how tightly two quantities move together along a straight line, from −1 to +1. Regression fits the best line through the cloud (by least squares, as in Linear Algebra, 34/04). Both are summaries, and any summary can hide the story. In 1973 the statistician Francis Anscombe made four data sets with, to two decimals, the same mean, the same spread, the same r (0.82) and the same line (y = 3 + 0.5x). Look at each.

  
    Interactive · Four data sets, one set of numbers
    
      Set 1Set 2Set 3Set 4
    
    
      
      
    
    Without JavaScript: all four Anscombe sets have mean x = 9, mean y = 7.50, correlation 0.816 and line y = 3.00 + 0.50x. Set 1 is a noisy line, set 2 a curve, set 3 a line with one outlier, and set 4 a single far point that creates the whole slope.
  
  The numbers agree and the stories do not: one noisy line, one curve, one line with a rogue point, one vertical stack held up by a single point. Always plot first.
  The other trap is the famous one. Ice-cream sales and drownings rise together, because summer heat drives both. A confounder (a hidden third factor) makes two things move together without either causing the other. The remedy graduates know is randomised assignment: if a coin decides who gets the treatment, nothing else can be systematically linked to it. Where you can't randomise, causal claims need much more than a line through a cloud.

  Engineering: tolerance stacks, reliability and Monte Carlo
  Five parts, each 10 mm ± 0.1 mm, line up in an assembly. What is the worst the total can be out? Worst-case stacking adds the tolerances: ±0.5 mm. It is safe and expensive, because it assumes every part is at its extreme in the same direction at once. If the parts vary independently, variances add, so the standard deviations add in quadrature: root-sum-square, √5 × 0.1 = ±0.224 mm. For five parts that is under half the worst-case figure.
  When the formulas get hard, engineers use Monte Carlo: let the computer build the assembly ten thousand times with random parts and count how often it fails. Here each part's tolerance is treated as three standard deviations. Choose how many parts are stacked.

  
    Interactive · Monte Carlo tolerance stack
    
      Parts in the stack
      Build 10,000 assemblies again
      
      
    
    Without JavaScript: five parts at ±0.1 mm (three standard deviations) give a worst-case limit of ±0.5 mm, a root-sum-square limit of ±0.224 mm, and in simulation only about 0.3% of assemblies exceed the root-sum-square limit; none come close to 0.5 mm.
  
  The honest caveat is that root-sum-square assumes the parts vary independently and are centred. A tool wearing in the same direction on every part breaks that assumption, and so does one supplier's batch. Engineers disagree about when to trust statistical stacking over worst-case; safety-critical designs often insist on worst-case.
  Reliability follows the same arithmetic. A system that needs all of its parts is a series system, and probabilities multiply: ten components that each work 99% of the time give 0.9910 ≈ 90.4% for the whole. Add a backup in parallel and failure needs both to fail: two units that each work 90% of the time give 1 − 0.1² = 99%. The two rules drive why aircraft duplicate systems and why complicated machines need very reliable components.

  What a graduate just sees
  
    Every number has a spread behind it; the first question is "how scattered?", not "what is it?".
    Spread, not the average, decides how many parts are scrapped, and centring matters too.
    Averages of many independent things are bell-shaped, whatever the source looked like.
    Halving the uncertainty costs four times the data.
    A 95% interval is a claim about a method: it misses one time in twenty by design.
    A p-value is "how surprising is this if nothing is going on", not "how likely is it true".
    Test twenty things and one will look significant; ask how many were tried.
    Plot the data; correlation is not causation, and only randomising breaks a confounder.
    Independent errors add in quadrature; series reliabilities multiply.
  

  Where it shows up next
  
    Machine Learning (34/32): training is fitting a model to noisy data; ideas of overfitting, error bars and Bayes are this course made bigger.
    Materials Science & Manufacturing (34/13): process capability and scrap rates are the normal curve against a spec limit.
    Mechanics of Materials (34/11) and Machine Design (34/19): safety factors exist because loads and strengths are distributions, not numbers.
    Systems Engineering (34/22): reliability budgets multiply series probabilities and add redundancy to push them back up.
    Computer Security (34/33) and Algorithms (34/24): good random numbers, randomised algorithms, and the odds of guessing a secret.
  

  If you want to go deeper
  
    📘 Douglas Montgomery and George Runger, Applied Statistics and Probability for Engineers (Wiley; 7th edition). The standard engineering course text, with plenty of tolerance, quality and reliability problems.
    📘 Joseph Blitzstein and Jessica Hwang, Introduction to Probability (CRC Press; 2nd edition, 2019). Best for deep intuition about probability itself, from Harvard's Stat 110.
    📖 David Spiegelhalter, The Art of Statistics (2019): the most readable route for learning alone, built around real cases and honest about uncertainty.
    🌐 MIT OpenCourseWare 18.05, Introduction to Probability and Statistics (Spring 2022): free notes and problem sets, with Bayesian and classical statistics side by side.
  

  
    The takeaway
    Nothing you measure is exact, so every number is a shape. Summarise a shape by its centre and its spread, trust averages more as the square root of the sample, read a confidence interval as a promise about a method, treat a p-value as a measure of surprise, and plot before you believe a correlation.
    Try this: next time you see a poll or a "significant" headline, find the sample size and ask how many things were tested. Then, in the dice widget above, add four dice and watch the tent turn into a bell.
  

  
    Sources & Further Reading
    Sources. The normal-curve rule (68.3%, 95.4%, 99.7%), the central limit theorem, confidence intervals, the binomial test, the Beta posterior, root-sum-square stacking and series/parallel reliability are standard textbook results (Montgomery and Runger; MIT 18.05); every number shown was recomputed in code. Wasserstein and Lazar, "The ASA's Statement on p-Values: Context, Process, and Purpose", The American Statistician (2016). Open Science Collaboration, "Estimating the reproducibility of psychological science", Science (2015): 97% of originals significant against 36% of replications. Anscombe, "Graphs in Statistical Analysis", The American Statistician (1973). Early forms of the central limit theorem: de Moivre (1733) and Laplace (c. 1810), as surveyed in recent histories such as Ulyanov, "From classical to modern central limit theorems" (arXiv 2405.19828).
    Go deeper. The books and course above; Wasserstein and Lazar's statement is free and short (doi.org/10.1080/00031305.2016.1154108).


Step 4. Accounting: Every Transaction Has Two Sides
---------------------------------------------------
Why: Marketing spends money; accounting shows whether it came back.
Accounting is the language a business uses to keep score. Learn one rule, that everything a firm owns is paid for by someone, and you can read how any company is really doing, including why a profitable one can run out of cash.

A bakery has its best month ever. Orders are up, the ovens never cool, the owner shows her bookkeeper a sales total with a happy number on it. Three weeks later she cannot make payroll. Nothing was stolen and nobody lied. The restaurants that bought her bread pay on 60-day terms, the flour supplier wants money in 14 days, and the new oven was paid for in cash on day one. Sales, profit and cash are three different things, and the business ran out of the one you cannot print.
  That gap is the first thing an accounting course teaches you to see. The second is that every number in a company's reports is connected to every other by one very old rule, small enough to write on a napkin. A graduate does not memorise rules about reports. They carry that napkin rule in their head, and the reports fall out of it.

  
    The course in one box
    
      Who takes itEvery business, accounting and finance student in the first or second year; engineers, lawyers and entrepreneurs often take a short version. At Wharton it sits in the undergraduate business core, and the first-year Required Curriculum at Harvard Business School opens with financial reporting and control. Programmes split it in two: financial accounting (reporting to outsiders) and managerial accounting (numbers for decisions inside the firm).
      What it is forRecording what a business owns, owes and earns, so that owners, lenders, tax authorities and managers can all read the same story.
      Comes afterNothing. It is a starting subject, though the arithmetic is ordinary school maths.
      Leads toCorporate Finance (34/51), then Entrepreneurship (34/56). The ideas also feed Operations Management (34/52) and Pricing (34/61).
      What this piece gives youYou will not be able to prepare a company's audited accounts. You will be able to post a transaction, see why the books always balance, and spot the difference between profit and cash. Easier first reads in the library: Where Money Actually Comes From (04/02) and Reading the Business You Work For (04/06). This one goes further and shows all three statements moving together. It describes how things work; it is not tax, legal or investment advice.
    
  

  Every transaction has two sides: the accounting equation
  Here is the napkin rule. Everything a business owns was paid for by someone: either by people it owes (lenders, suppliers) or by the owners. There is no third source. So:
  Assets = Liabilities + Equitywhat the firm owns = what it owes + what the owners have put in or left in
  Because this is always true, every transaction must touch the books in at least two places, so that the equation stays balanced. That is double-entry bookkeeping. The first detailed printed description came from the friar Luca Pacioli, in a Venice maths textbook of 1494, but he codified a method that Italian merchants had used for two centuries before. A debit and a credit are just the left and right sides of a two-column ledger; the words carry no moral meaning. Assets and expenses grow on the debit side, liabilities, equity and revenue grow on the credit side, and total debits always equal total credits.
  Below is a small start-up, posted one transaction at a time. Watch three things: the equation staying balanced, profit appearing on the income statement, and the cash flow statement telling a different story.
  
    Interactive · Post a year of transactions
    
      Post next
      Post all
      Reset
    
    
    
      Assets$Liabilities and equity$
      CashSupplier owed
      Customers owe usBank loan
      InventoryOwner's capital
      Equipment (net)Profit kept
      Total assetsTotal liabilities + equity
    
    
      Income statement$Cash flow statement$
      RevenueOperations
      Cost of goods soldInvesting
      RentFinancing
      DepreciationChange in cash
      ProfitCash from operations
    
    
    An invented start-up, for illustration. No tax, no interest, to keep the arithmetic visible.
  
  Post all nine and you land on a result worth remembering: profit of $1,700, but cash from operations of −$800. The firm has $8,200 in the bank, but only because the owner and the bank put in $15,000 and $6,000 went on equipment. The sale was booked the moment the goods left, yet only $4,000 of the $6,000 had come in. And the $1,000 of depreciation reduced profit without a single dollar leaving. Three statements, one set of events.

  Profit is not cash: income statement against cash flow
  The income statement (or profit and loss) answers: over this period, how much did we earn and spend? It runs from revenue down to profit. The cash flow statement answers a different question: how much money actually moved in and out of the bank? It is split into operations (the day-to-day business), investing (buying and selling long-lived assets) and financing (loans and owners' money). A firm can be fine on one and dying on the other.
  The usual trap is a growing business that sells on credit. It pays for materials and staff now and is paid back later, so every extra sale makes the cash hole deeper before it fills it. Try it: a business makes a 30% gross margin, spends $2,000 a month on overheads, starts with $10,000 and sells $10,000 in month one. Change how fast it grows and how long customers take to pay.
  
    Interactive · Profitable and broke
    Sales growth per month
      
    Days customers take to pay
      
    
    
    Illustrative business: month-one sales $10,000, goods cost 70% of sales and are paid at once, overheads $2,000 a month, opening cash $10,000. The line is cash at each month end.
  
  At 10% growth and 60 days, the business earns about $40,150 of profit over the year and still ends with −$4,300 in the bank, dipping to about −$9,600 in month 5. Set the days to zero and it ends with $50,150 in cash. The same sales, the same profit; only the waiting changed. This is why lenders watch cash flow and why fast-growing firms fail: growth eats cash first.

  The balance sheet is a photograph, not a film
  The income statement covers a stretch of time. The balance sheet is a snapshot of one day: what the firm owns on the left, what it owes and who funded it on the right. Same equation, frozen. Profit flows in as "profit kept" (retained earnings) at the bottom, which is how the three statements lock together: the income statement's profit raises equity on the balance sheet, and the cash flow statement explains the change in the cash line.
  
    StatementTimeQuestion it answers
    Balance sheetone dayWhat do we own and owe, and who funded it?
    Income statementa periodDid we earn more than we spent?
    Cash flow statementa periodWhere did the cash come from and go?
  
  A graduate reads a balance sheet like a map. Lots of receivables and inventory and little cash? Growth is tying up money. Debt far bigger than equity? Small bumps in earnings will hurt. Equipment large and old? Replacement is coming. You do not need the numbers to be exciting; you need them to be consistent.

  Accrual, cash and depreciation: matching cost to use
  There are two ways to decide when something counts. Cash accounting records a sale when the money arrives and a cost when it is paid. Small shops and many individuals use it, and some tax systems allow it. Accrual accounting records revenue when it is earned and costs when they are incurred, whatever the bank says. Large companies are required to use accrual because it matches effort with reward: if you bought a delivery van for $12,000, it would be silly to say your van business lost $12,000 in the first year and made a fortune in the next four.
  That is the logic of depreciation: spread the cost of a long-lived asset over the years it helps earn money. The simplest way is straight line: (cost − expected scrap value) ÷ years of life. Faster methods, like declining balance, write off more early and less later. Depreciation is the classic non-cash expense: it lowers profit and tax, and no money moves that year. The cash left when the asset was bought.
  
    Interactive · Spread the cost of a van
    
      Straight line
      Double declining
    
    Purchase cost
      
    Useful life
      
    Scrap value at the end
      
    
    
  
  With a $12,000 van, $2,000 scrap value and five years, straight line charges $2,000 a year. Double declining charges $4,800, $2,880, $1,728, then only $592 in year 4 because the book value cannot go below scrap, and nothing in year 5. Same total, different timing. Tax rules often dictate which method a business may use; the accounting question is which one best shows the van wearing out.

  Reading ratios: five numbers that tell a story
  A statement is a list of numbers. A ratio turns two of them into a judgement. Accountants and lenders lean on a short set, each one answering a plain question: margin (how much of each sale is left as profit?), current ratio (can we pay the bills due within a year?), days sales outstanding (how long do customers take to pay?), debt to equity (how much of the funding is borrowed?), and return on equity (what did the owners' money earn?).
  Current ratio = current assets ÷ current liabilitiesa ratio under 1 means the bills due soon are bigger than the cash and near-cash to pay them
  Ratios are only meaningful against something: the firm's own past, or other firms in the same business. A supermarket's 2% margin and a software firm's 70% margin are both fine. Here is a small company. Move its numbers and watch the ratios and the verdict react.
  
    Interactive · Read a small company
    Sales for the year
      
    Cost of goods as share of sales
      
    Other running costs
      
    Customers owe us
      
    Cash in the bank
      
    Bills owed to suppliers
      
    Bank debt (8% interest)
      
    
    Invented figures in $ thousands. Inventory is fixed at 90 and equipment at 300; equity is whatever is left after liabilities. No tax.
  
  With the starting numbers the firm earns about $104,000, a net margin of 13%, a current ratio of 3.5, customers who take 59 days to pay, debt of 0.67 times equity and a return on equity near 35%. Now push the debt to $400,000 with the same assets. Equity shrinks to $100,000, debt to equity jumps to 4.0, interest cover falls from 7.5 to 3.8 times, and return on equity leaps to 88%. The last number looks wonderful and is the dangerous one: a thin base of owners' money flatters every return. No single ratio decides; the pattern does.

  Managerial accounting: fixed, variable and the break-even point
  Everything so far is for outsiders. Managerial accounting is for decisions: should we make this, price that, close this shop? It starts by splitting costs into two kinds. Variable costs rise with each unit you make (flour, cups, shipping). Fixed costs do not (rent, salaries, insurance). Price minus variable cost is the contribution margin: what each sale contributes toward covering the fixed costs and, after that, towards profit.
  Break-even units = fixed costs ÷ (price − variable cost per unit)below this volume you lose money; above it, each extra sale adds its contribution margin to profit
  A coffee cart sells a cup at $4.00 and spends $1.50 on beans, milk and cup, so each cup contributes $2.50. With $3,000 a month in rent and wages it needs 1,200 cups to break even. At 1,500 cups it makes $750; the gap between 1,500 and 1,200, a fifth of sales, is its margin of safety. Cut the price a little and see how much more volume you need.
  
    Interactive · Break-even for a coffee cart
    Price per cup
      
    Variable cost per cup
      
    Fixed costs per month
      
    Cups sold per month
      
    
    
  
  The shape is what graduates keep: a business with high fixed costs and a fat margin per unit is a lever. Below break-even it loses a lot; above it, profit rises fast. An airline with half-empty planes and a software firm with no cost per extra customer both live on this curve. It is also the first tool of Entrepreneurship (34/56).

  Two rulebooks, and what happens when the books lie
  Someone has to decide how the rules apply. Most of the world uses IFRS, the standards of the International Accounting Standards Board; the United States uses its own US GAAP, written by the FASB. In the IFRS Foundation's 2018 overview, 144 of 166 profiled jurisdictions required IFRS for their publicly accountable companies, including 15 of the 20 G20 economies; the US lets foreign companies list there using IFRS but requires domestic ones to use US GAAP. A 2002 agreement in Norwalk, Connecticut began a decade of work to converge the two, and many differences remain.
  
    QuestionIFRSUS GAAP
    StylePrinciples-based; more room for judgementMore detailed rules
    LIFO inventory (last in, first out)Not allowedAllowed
    Written-down inventory later recoversReversal allowedReversal not allowed
    Development costsCan be capitalised once criteria are metGenerally expensed
  
  The table is a sample, not a full comparison, and standards change; check the current text before relying on any line. Neither rulebook is "better": each reflects choices, and a firm's profit can differ under the two for exactly the same activity. An auditor, an independent accounting firm, tests whether the statements follow the rules; it does not guarantee the numbers are true.
  What happens when the rules are bent shows why. Enron was a real company with real pipelines and a huge energy-trading business. Its accounts used mark-to-market accounting (the SEC approved it for the firm's gas contracts in 1992), which let it book the estimated profit on a long-term contract on day one, based on its own forecasts. It also used special purpose entities, partnerships kept off its balance sheet, to hide debt and losses. A rule allowed a vehicle to stay unconsolidated if outside investors put in at least 3% of its capital; the later investigation by a board committee (the Powers Report) found that some failed even that. Chief financial officer Andrew Fastow ran two partnerships, LJM1 and LJM2, in which he profited personally, and which traded with Enron itself.
  The stock, around $90 in August 2000, was below $1 by the end of November 2001. Enron restated years of results, then filed for bankruptcy on 2 December 2001 with $63.4 billion in assets, the largest US bankruptcy at the time. Former chief executive Jeffrey Skilling was convicted in 2006 on 19 of 28 counts and Kenneth Lay was convicted but died before sentencing. Auditor Arthur Andersen was convicted of obstruction in June 2002 for shredding documents; the Supreme Court unanimously overturned that in 2005 because of the jury instructions, but by then the firm was gone. Congress passed the Sarbanes–Oxley Act that July. The lesson is not "accountants are crooks" but that rules, incentives and oversight all need to hold, and that a company's reported profit is an opinion built on choices.

  What a graduate just sees
  
    Everything owned was paid for by someone, so the books always balance: assets = liabilities + equity.
    Every transaction has two sides; if only one side shows up, something is missing.
    Profit and cash are different. A growing firm that sells on credit can be profitable and bankrupt.
    The balance sheet is a photograph; the income and cash statements are films between two photographs.
    Depreciation is a real cost spread over years, but it is not cash leaving.
    A ratio means nothing alone; compare it with the firm's past and its competitors.
    Fixed costs make a business a lever: above break-even each sale is mostly profit.
    Reported earnings depend on choices. Always ask what the accounting rule allowed.
  

  Where it shows up next
  
    Corporate Finance (34/51): values a firm from the cash flows that accounting reports, so cash is the language.
    Entrepreneurship (34/56): break-even, burn and runway are managerial accounting applied to a start-up.
    Operations Management (34/52) and Pricing (34/61): cost behaviour and contribution margin decide what to make and what to charge.
    In the world: a loan application, a tax return, a job at any company, reading a news story about "profit warnings", or noticing that a business you love is quietly running on credit.
  

  If you want to go deeper
  
    📘 Textbook. Jerry Weygandt, Paul Kimmel and Donald Kieso, Financial Accounting (Wiley). A very widely assigned first-year text with plenty of practice problems.
    📘 Textbook. Ray Garrison, Eric Noreen and Peter Brewer, Managerial Accounting (McGraw Hill). The standard on cost behaviour and break-even; use the current edition.
    📘 Gentler. Thomas Ittelson, Financial Statements: A Step-by-Step Guide to Understanding and Creating Financial Reports. Plain English, built on small examples.
    🌐 Free course. The free OpenStax textbooks Principles of Accounting, Volume 1: Financial Accounting and its managerial companion volume. Free, peer-reviewed, with worked examples; the quickest route to a full first-year course at no cost.
  

  
    The takeaway
    Everything a firm owns was paid for by someone, so each transaction has two sides and the books balance. Three statements tell one story from three angles: the balance sheet is where you are, the income statement says what you earned, and the cash flow statement says what you can spend. Profit is an opinion built on rules; cash is a fact.
    Try this: pick a company you like, find its latest annual report, and read only three lines: net profit, cash from operations and cash at the end. If the first two disagree for several years, find out why.
  

  
    Sources & Further Reading
    Sources. Double entry and Pacioli's Summa de arithmetica (Venice, 1494): codified, not invented, per standard accounting histories. Standards: IAS 2 (LIFO prohibited, inventory reversals), IAS 38 (development costs) versus US GAAP; check current text. IFRS adoption: IFRS Foundation, Use of IFRS Standards around the world (2018 overview: 144 of 166 jurisdictions, 15 of 20 G20); the Foundation's profile page now lists 170 jurisdictions. The 2002 Norwalk Agreement: FASB and IASB memorandum. Enron: Wikipedia summary checked against the dated facts for the SEC's January 1992 mark-to-market approval, the stock peak and collapse, the 2 December 2001 Chapter 11 filing, the Skilling conviction on 25 May 2006, Lay's death on 5 July 2006, Andersen's conviction (15 June 2002) and reversal (Arthur Andersen LLP v. United States, 2005), and the Sarbanes–Oxley Act (30 July 2002); the 3% outside-equity rule and the Powers Report (2002) are from the board committee's report and secondary accounts. All interactive numbers were computed in node: the nine-step ledger ends at profit $1,700, operating cash −$800, cash $8,200, assets $16,700; the cash trap at 10% growth, 60 days ends at cash −$4,316 with profit $40,153; depreciation $2,000 a year straight line, $4,800 / $2,880 / $1,728 / $592 / $0 declining; break-even 1,200 cups, profit $750 at 1,500.
    
      🌐 OpenStax, Principles of Accounting, Volume 1
      🌐 IFRS Foundation: use of IFRS Standards by jurisdiction
      📘 Garrison, Noreen & Brewer, Managerial Accounting
      📘 Ittelson, Financial Statements


Step 5. Consumer Behavior: How People Notice, Want, Choose and Regret
---------------------------------------------------------------------
Why: How people notice, want, choose and regret, and what research says and doesn't.
Nobody weighs every option. People notice a sliver of what is in front of them, follow each other more than they admit, and regret in predictable ways. A graduate also knows which famous "findings" have held up and which were never real.

You walk into a supermarket for milk and walk out with milk, a chocolate bar and a magazine. Ask yourself afterwards what the shelf looked like and you could not draw it. You did not read forty labels. You did what everyone does: looked at a fraction of what was there, went to the familiar, glanced at what other people were taking, and decided in a few seconds.
  Consumer behavior is the study of how that actually happens, as opposed to how a spreadsheet says it should. It sits where marketing meets psychology. Its famous results fill popular books, and a good share of them have not survived careful checking. Part of what a graduate learns is how to tell which is which.

  
    The course in one box
    
      Who takes itMarketing majors, usually in year 2, straight after the introductory marketing course (it is the second course in the Wharton marketing concentration). Psychology, design and business students take it as an elective.
      What it is forPredicting what real people notice, prefer and do, and what they will regret, so that products, shelves and messages fit them.
      Comes afterMarketing Management (34/57) and Behavioral Economics (34/46).
      Leads toBrand Management (34/60), Pricing (34/61), and, through Marketing Management, Marketing Research & Analytics (34/59).
      What this piece gives youYou won't be able to run a consumer study. You will know how little a shopper takes in, why crowds amplify luck, and which famous claims to doubt. Easier first reads: The Biases That Fool Everyone (17/02). This one goes further and checks the claims.
    
  

  You notice far less than you think
  Sharp vision covers only a small patch at the centre of your gaze. The rest of the scene is filled in by expectation. That is why people can look straight at something and not see it. In the best-known demonstration, Daniel Simons and Christopher Chabris asked viewers to count basketball passes in a video; about half of them did not notice a person in a gorilla suit walk through the middle (Perception, 1999). The effect has been repeated many times. The lesson for marketing is blunt: attention is the scarce resource, and the shopper is not looking where you hope.
  A cousin is change blindness. If a change happens during a blink, a cut or a flicker, people miss large differences. Ronald Rensink and colleagues built a "flicker" method in 1997 that shows how hard it is to spot a change when a blank screen separates two versions. Try it: one item on the shelf changes colour. Find it.
  
    Interactive · Spot the change on the shelf
    
      Start a new round
      Blank between frames: on
      Show me
    
    
    Press Start. Tap the item you think is changing. The shelf flickers about twice a second.
    Illustrative demo. Frames alternate: original, (blank), changed, (blank). With the blank off, the change makes a visible jump and is much easier to see.
  
  With the blank on, most people take several cycles; with it off, the moving colour jumps out. A real shelf has hundreds of items and no obvious blank, but every blink, glance and pause plays that role. For brands this means a distinctive colour or shape is not decoration: it is how the product gets noticed at all (see Brand Management, 34/60).

  Learning to want, and the decision journey
  People are not born wanting a brand. They learn to. Three mechanisms matter. Repetition: Robert Zajonc showed in 1968 that merely seeing something again makes people like it a bit more, and the "mere exposure" effect has held up in a large meta-analysis by Robert Bornstein (1989), with the effect modest and fading if repetition is overdone. Association: pair a brand with a feeling people already like (a sunny beach, a trusted face) and the feeling rubs off, a form of conditioning. Memory: you can only choose what comes to mind. Researchers call the handful of options a shopper actually thinks about the consideration set.
  Then comes the decision journey. The textbook picture is a funnel; in 2009 McKinsey consultants (Court, Elzinga, Mulder and Vetvik) argued it is a loop with four moments: initial consideration, active evaluation, purchase, and the experience after, which feeds back into the next consideration. For many everyday goods the journey is nearly empty: you grab the usual. The more a purchase costs, or the more risky it is, the longer and more deliberate the journey becomes. A graduate asks first how much of a journey this product has at all.
  
    MomentWhat happensWhat it implies
    ConsiderationA few options come to mindIf you aren't recalled, you aren't chosen
    EvaluationComparing, asking, searchingReviews and friends matter most here
    PurchaseOften fast, at the shelf or screenAvailability and defaults decide
    AfterUse, satisfaction or doubtFeeds the next round
  

  Choice: is more always worse?
  The most-quoted study in the subject is the "jam study" of Sheena Iyengar and Mark Lepper (2000). In a supermarket, a table with 24 jams drew more people to stop than a table with 6, but about 3% of the people at the big table bought, against about 30% at the small one. The conclusion, that too much choice paralyses, became a bestseller idea (Barry Schwartz, The Paradox of Choice, 2004).
  Then researchers tried to repeat it. In 2010 Benjamin Scheibehenne, Rainer Greifeneder and Peter Todd pooled 50 experiments and found that the average effect of more options was close to zero, with large differences between studies. Alexander Chernev and colleagues (2015) ran their own meta-analysis and argued the effect is real under conditions: when the options are hard to compare, when the shopper has no clear preference, and when the set is large and complex. So the honest reading: choice overload happens, but it is not a law, and the famous 3% versus 30% was not a typical size.
  A simple model shows why both can be true. Suppose every option has a random value for you between 0 and 1. The best of n options has expected value n ÷ (n + 1): more options always raise the best you could find, with diminishing returns. But looking costs effort. Subtract a small cost per option and a sweet spot appears.
  net = n ÷ (n + 1) − c × nn = options · c = effort cost of each option as a share of full satisfaction
  
    Interactive · How many options is best?
    Cost of looking at each option
      
    Options on the shelf
      
    
    
    A toy model, not data. It shows the shape of the argument: more options help up to a point, effort costs rise linearly, and the peak depends on how costly looking is for this person and this product.
  
  At half a point per option the best number is 13, but the curve is flat near the top, so a shelf with 6 or with 24 is within a few points of 13. The model does not say the jam study was wrong. It says the effect depends on effort, which is why big, confusing sets can hurt while familiar categories (milk, toothpaste) usually do not.

  Other people decide with us: reference groups and social proof
  A reference group is anyone whose view you use as a yardstick: family, colleagues, strangers who look like you. Robert Cialdini called the habit of using other people's behaviour as evidence social proof (Influence, 1984). It is sensible: if a restaurant is full, it is probably good. The risk is that proof can be circular, with each person copying the last.
  Matthew Salganik, Peter Dodds and Duncan Watts tested this in 2006. 14,341 participants downloaded unknown songs from an online "music market". Some saw others' download counts; some did not. When people could see the counts, the popular songs got much more popular, the unpopular ones less, and which songs won differed from one run of the experiment to the next. Quality mattered a little (the best songs rarely did badly and the worst rarely did well), but luck early on was amplified. The next model captures the mechanism: 200 people choose in turn between two equally good options.
  
    Interactive · Twenty parallel worlds of 200 choosers
    How much people copy the crowd so far
      
    Run again
    
    
    A toy model in the spirit of the 2006 study (not its data). Each person picks A with probability (1 − w) × ½ + w × the share of A so far. Both options are identical; each dot is one world's final share of A.
  
  With no copying, every world ends near 50-50. Turn the influence up and the worlds spread out: in some, A takes over; in others, B, though the two options are identical. Success looks like quality from inside any single world. This is why marketers pay for early reviews, why "bestseller" labels work, and why a hit can be partly an accident.

  The same ad, a different country
  Consumers carry a culture. Geert Hofstede's 1980 study, based on IBM employee surveys in more than 50 countries (about 117,000 matched samples, 1967–73), grouped national differences along dimensions such as individualism versus collectivism, power distance and uncertainty avoidance. In his published scores, the United States rates about 91 on individualism, while countries like Hong Kong sit near 25. Marketers used this to guess whether an ad should talk about "you" or about "family".
  Treat it as a hypothesis, not a map. Critics such as Brendan McSweeney (2002) challenge the method: the sample was one company and mostly men, countries are not uniform, a national average says little about any individual, and the data is old. A useful rule: use culture to ask a question ("would this joke, colour or family role mean the same here?") and then test it with local customers. Many tales about product names that "meant something rude" abroad are urban legends, so check before repeating one.

  Famous claims, checked
  The field has a pile of results that everyone has heard. Some were real, some shrank, some were never what they were said to be. Guess first; then see.
  
    Interactive · Held up, shrank, or never real?
    Claim 1 of 7
    
    
      Held up
      Shrank or depends
      Not real as told
    
    
    Next claim
  
  Notice how often the number in the story is the part that fails, while a smaller, real effect remains. Two lessons for working marketers: ask for the sample and whether anyone repeated it, and be wary of any single "X% of decisions" figure.

  After the purchase: regret and doubt
  Leon Festinger's cognitive dissonance (1957) explains why people, after a costly purchase, seek reasons they chose well and avoid evidence they didn't: the mind dislikes holding "I chose it" and "it was a mistake" together. Sellers use this with follow-up messages that confirm the choice. Regret has a time pattern too. Thomas Gilovich and Victoria Medvec (1995) found that in the short run people regret what they did, but looking back over a long time they regret what they did not do.
  On the shelf, regret and choice interact. Barry Schwartz and colleagues (2002) found that "maximisers", who check everything for the best, report less satisfaction and more regret than "satisficers", who take what is good enough; the work is correlational and later studies are mixed, so treat it as a tendency, not a verdict. Returns policies, free trials and reviews are all ways of lowering the price of regret.

  What a graduate just sees
  
    The shopper looks at very little; the first job of a pack or a page is to be noticed.
    Most everyday buying is quick and habitual; the "journey" is short for cheap, familiar things.
    Repetition and association build liking before any argument is heard.
    More choice is not automatically worse; it depends on effort and on whether people know what they want.
    Crowds amplify luck: popularity is partly a story about earlier popularity.
    Culture is a question to test locally, not a stereotype to assume.
    When a study has a dramatic number, check whether it was repeated.
    People regret, and sellers can lower that cost by making returns and reassurance easy.
  

  Where it shows up next
  
    Brand Management (34/60): mental availability is the consideration set, and distinctive assets are the answer to attention.
    Pricing (34/61): how people perceive a price, from anchors to "is this fair".
    Marketing Research & Analytics (34/59): how to ask, test and check whether a finding replicates.
    Behavioral Economics (34/46): the formal sister of this subject, with the same replication debates.
    In the world: why retail shelves are laid out as they are, why apps rank "most popular", and why returns policies exist.
  

  If you want to go deeper
  
    📘 Textbook. Michael Solomon, Consumer Behavior: Buying, Having, and Being (Pearson). The standard undergraduate text, with many applied cases.
    📘 Textbook. Wayne Hoyer, Deborah MacInnis and Rik Pieters, Consumer Behavior (Cengage). More research-focused.
    📘 Gentler. Robert Cialdini, Influence (1984; revised editions). Readable and full of examples. Daniel Kahneman, Thinking, Fast and Slow (2011) is the companion, though read it knowing some of its cited priming studies have been questioned.
    🌐 Free course. MIT OpenCourseWare 15.810 Marketing Management covers the consumer side in its middle sessions; there is no single free consumer-behavior course I could confirm, so pair it with the books.
  

  
    The takeaway
    People notice little, copy others, decide fast and regret in patterns. Design for the sliver of attention you will get, remember that popularity feeds itself, and check any famous claim for a replication before relying on it.
    Try this: next time you shop, notice the first three things you looked at and the last thing you picked. Then ask which was seen first, and which came from habit.
  
  
    Sources & Further Reading
    Sources. Simons & Chabris, "Gorillas in our midst", Perception (1999). Rensink, O'Regan & Clark (1997), flicker method. Zajonc (1968); Bornstein (1989). Court et al., "The consumer decision journey", McKinsey Quarterly (2009). Iyengar & Lepper, JPSP (2000); Scheibehenne, Greifeneder & Todd, Journal of Consumer Research (2010: 50 experiments, mean effect near zero); Chernev, Böckenholt & Goodman, Journal of Consumer Psychology (2015). Salganik, Dodds & Watts, Science (2006; 14,341 participants). Hofstede (1980; 117,000 IBM samples, via Wikipedia summary); McSweeney (2002). James Vicary's 1957 popcorn claim (57.5% and 18.1% increases) was called "a gimmick" by Vicary in a 1962 Advertising Age interview. Festinger (1957); Gilovich & Medvec (1995); Schwartz et al. (2002). Maslow's pyramid: Bridgman, Cummings & Ballard, "Who built Maslow's pyramid?", Academy of Management Learning & Education (2019). The simulations are toy models; the numbers (best n = 13 at 0.5% cost; herding spread) were computed in node. This is not business advice.
    
      🌐 MIT OCW 15.810 Marketing Management
      📘 Solomon, Consumer Behavior
      📘 Cialdini, Influence


Step 6. Behavioral Economics: How People Actually Choose
--------------------------------------------------------
Why: The biases behind defaults, anchors and "only 3 left".
Standard economics assumes calm, consistent, self-interested calculators. Behavioral economics measures how real people depart from that, and, just as important, which of its famous findings have survived being checked.

You sign up for a gym at 8 a.m. in January, full of resolve, and pay monthly for a year. You go four times. A shop shows a jacket "was £200, now £120" and it suddenly feels like a bargain, though you had no idea what it was worth a minute ago. A company enrolls new staff in a pension plan automatically, and far more people end up saving, although nothing about their finances changed.
  None of this is irrational in the sense of being random. It is irrational in a predictable way, and predictable is what an economist can use. Behavioral economics collects those regular departures from the textbook chooser and builds them into models. It is also a field that has been through a hard decade, in which many headline results were re-tested and some did not survive. This piece teaches both: the ideas, and the report card.

  
    The course in one box
    
      Who takes itEconomics majors, usually in year 3 after microeconomics and statistics; marketing, psychology and business-school programmes teach it as consumer behaviour or decision-making. MIT, for example, offers a course called Psychology and Economics (14.13).
      What it is forPredicting where real choices depart from the rational-agent benchmark, and designing prices, defaults and policies that work for real people.
      Comes afterMicroeconomics (34/42) and Probability & Statistics (34/05).
      Leads toConsumer Behavior (34/58), and it feeds Pricing (34/61) and Marketing Management (34/57). Econometrics (34/44) is the toolkit for judging which of its findings are real.
      What this piece gives youYou will not be able to design a behavioural experiment. You will know the main effects, roughly how big they are, and which classics were overturned. Easier first reads: The Biases That Fool Everyone (17/02), How to Make a Hard Decision (17/05) and Why You Understand Less Than You Think (17/10). This piece goes further by asking how much each bias is worth and whether it held up.
    
  

  Bounded rationality: shortcuts that usually work
  The textbook consumer weighs every option, knows her own tastes and never makes arithmetic mistakes. Herbert Simon argued in the 1950s that this ignores the obvious: minds are limited, time is short, information is incomplete. People do not optimise, they satisfice: search until something is good enough, then stop. He called this bounded rationality. Daniel Kahneman and Amos Tversky later catalogued the rules of thumb, or heuristics, that people use, and the systematic errors each one produces. Judging how likely something is by how easily examples come to mind (availability), by how much it resembles a stereotype (representativeness), or by starting from a number you were handed (anchoring), all work most of the time, and all fail in patterned ways. Kahneman received the 2002 economics Nobel prize for this work.
  A worked example, from the "cognitive reflection" literature: a bat and a ball cost £1.10 together; the bat costs £1.00 more than the ball. How much is the ball? The quick answer is 10p. Check it: ball 10p, bat £1.10, total £1.20. If the ball costs x, the bat costs x + 1.00, so 2x + 1.00 = 1.10 and x = 5p. The wrong answer comes fast and feels right; the right one takes a second look. Biases are mostly this: a fast process that is usually fine, with no alarm bell when it is not. The rest of the piece is a tour of the best-studied ones, with a growing sense of how much to trust each.

  Prospect theory: gains, losses and the kink
  Kahneman and Tversky's 1979 paper in Econometrica proposed prospect theory as a description of how people value risky outcomes. Three ingredients. First, people judge outcomes as gains or losses from a reference point (usually where they are now), not as totals of wealth. Second, diminishing sensitivity: the difference between winning 0 and 100 feels larger than between winning 1,000 and 1,100. Third, loss aversion: a loss hurts more than the same gain pleases. Their 1992 estimates were roughly α ≈ 0.88 for curvature and λ ≈ 2.25 for loss aversion.
  v(x) = xα for gains, v(x) = −λ·(−x)α for lossesλ above 1 makes the curve steeper for losses; α below 1 flattens it as amounts grow
  
    Interactive · Would you take the coin flip?
    Loss aversion λ
      
    Curvature α
      
    Win if heads (lose 100 if tails)
      
    
    
    A simplified prospect-theory value function, ignoring probability weighting. The values are in arbitrary "felt" units. At λ = 1 and α = 1 the chooser is the textbook risk-neutral one.
  
  Without JavaScript: with λ = 2.25 and α = 0.88, a 50:50 chance to lose 100 is only worth taking if the win is at least about 251; with λ = 2 and α = 1 you need exactly 200.
  At the default settings a coin flip that risks 100 and offers 150 has a positive average (+25 in money) but a negative felt value. Slide the win up to about 250 and it flips. People often refuse such bets, and prospect theory says why. How robust is loss aversion itself? That is an unfinished argument, taken up at the end.

  Anchoring and framing: the same facts, different pull
  In the best-known anchoring experiment, Tversky and Kahneman (1974) spun a wheel that stopped on 10 or 65, then asked people what percentage of United Nations members are African countries. People who saw 10 gave a typical answer near 25%; those who saw 65 answered near 45%. A random number pulled the estimate. The sliding tool below interpolates between those two reported medians (an illustration of the size of the pull, not data beyond them).
  
    Interactive · The anchor's pull
    Number the wheel showed
      
    
    
    Line through the two reported median estimates (anchor 10 gives 25; anchor 65 gives 45). Between them, every 1 point of anchor moves the typical estimate by about 0.36 points.
  
  The anchoring index is the shift in estimates divided by the shift in anchors: (45 − 25) ÷ (65 − 10) ≈ 0.36. Anchoring is one of the sturdier findings in the field, though a few studies with obviously random anchors have not found it. Framing is the cousin: Tversky and Kahneman's 1981 "Asian disease" problem describes a programme for 600 people at risk either as "200 will be saved" or as "400 will die", and both describe the same outcome. In the original, a majority took the sure option when it was framed as lives saved and a majority took the gamble when framed as deaths. Framing effects have been re-tested widely, and the direction usually holds; the size depends heavily on the wording and the people asked.

  Present bias: today versus later
  Standard theory discounts the future at a steady rate, so your preference between "£100 on day d" and "£110 on day d + 7" stays the same as the days pass. Real people do not. Richard Thaler found in 1981 that people would be indifferent between $15 now and $30 in three months, $60 in a year or $100 in three years, implying discount rates that fall steeply with the delay. George Ainslie found many people preferred $50 now to $100 in six months but $100 in nine months to $50 in three: the same choice, six months further away, reverses. David Laibson (1997) captured this in the quasi-hyperbolic model: everything after today gets an extra one-off discount β below 1.
  value of reward x at delay t = β · δt · x for t > 0, and x if nowδ is ordinary patience per day; β below 1 is "present bias"
  
    Interactive · Plan today, choose tomorrow
    Present bias β (1 = none)
      
    Days until the smaller reward
      
    
    Choice: £100 on that day, or £110 a week later. Patience δ = 0.995 per day (about 3.4% lost per week). Invented parameters for illustration.
  
  With β = 0.7 your plan today is to wait for the £110 whenever the smaller prize is a day or more away. When the day arrives and £100 is available now, you take it, against your own plan. That plan-then-break pattern is the gym membership, the unfinished diet, the unopened savings account. People who know they are like this buy commitment devices: a locked savings account, a bet with a friend, deleting the app. Set β = 1 and the reversal vanishes.

  Defaults and nudges: design the easy path
  If people are inattentive and slightly present-biased, whatever happens when they do nothing matters enormously. Madrian and Shea (2001) studied a large US company that switched its 401(k) pension plan to automatic enrolment, with the plan's economic terms unchanged: participation was significantly higher under automatic enrolment, and many employees stayed at the default contribution rate and fund. National organ-donation rates tell a similar story: about 12% consent in opt-in Germany against 99.98% in opt-out Austria is the oft-cited contrast, though it mixes in other causes, and Spain, with the highest donor rate, credits its transplant coordination more than its law.
  
    Interactive · Opt-in or opt-out?
    Share of people who actively decide
      
    Share who truly want to join
      
    
    
    A toy model: active deciders follow their own wish; everyone else stays at the default. Real take-up depends on many more things.
  
  The toy model shows the logic: the less attentive the population, the more the default decides, and the fewer people are choosing at all. This is the idea behind the nudge, a term Richard Thaler and Cass Sunstein made famous in their 2008 book: change the "choice architecture" without forbidding anything. It is also the cue for a warning. Whether nudges work at scale is disputed (report card below), and the ethics of steering people are argued too: critics worry about manipulation, supporters reply that every choice has some default, so someone has to set it.

  Social preferences: the ultimatum game
  Standard theory predicts people care only about their own payoff. In the ultimatum game (Güth, Schmittberger and Schwarze, 1982), a proposer splits 100 with a responder who can accept or reject; on rejection both get nothing. A pure money-maximiser would accept 1, so the proposer would offer 1. Real responders do not. Across 37 studies, proposers typically offered about 40%, and offers below about 30% were often rejected, so responders pay to punish what they see as unfair. Fifteen small-scale societies studied by Henrich and colleagues showed that what counts as fair differs across cultures, so the "typical" figures belong to the places where most experiments have been run.
  
    Interactive · Make the offer
    Your offer (out of 100)
      
    Typical minimum responders accept
      
    How much responders differ
      
    
    Each responder has a private minimum, spread as a bell curve around the typical value. Invented parameters: the point is the shape of the trade-off, not a forecast.
  
  Set the minimum and the spread to near zero (a textbook responder): the best offer collapses to about 3. Put them near what the studies describe and the best offer jumps to about 39, because a stingy offer risks everything. A proposer who offers fairly may be motivated by fairness or by fear of rejection, and the game alone cannot say which.

  What held up, what shrank, what failed
  Behavioral economics grew out of psychology, and psychology went through a replication crisis. In 2015 the Open Science Collaboration re-ran 100 published psychology studies and found about 36% replicated successfully by their main criterion. Economics did better but not perfectly: a 2016 project re-ran 18 laboratory experiments from top economics journals, and 11 (about 61%) replicated; a 2018 project re-ran 21 social-science papers from Nature and Science, and 13 (about 62%) replicated, typically with smaller effects. The report card below sorts the findings used in this piece and a few famous neighbours. Use the buttons to filter it.
  
    Interactive · The report card
    
      All
      Held up
      Shrank
      Failed
      Disputed
    
    
    
  
  A few lessons. Directions hold more often than sizes: defaults, anchors and present bias point the way the theory says, but their effects in the field are often smaller than the first lab study. Big, flashy, one-off findings are the riskiest: ego depletion and social priming were built on small samples and clever stories. Not every "bias" is one idea: loss aversion, the endowment effect and the status quo bias may be one force or three, and experts differ. And published averages exaggerate, for the reason shown in Econometrics (34/44): significant results survive the journal filter, and lucky overestimates are the ones that pass it.

  What a graduate just sees
  
    People use shortcuts that usually work and sometimes fail in patterned ways; the pattern is what makes it a science.
    Outcomes are judged from a reference point, and losses weigh more than equal gains, probably about twice, but the figure depends on context.
    A number you are handed pulls your estimate toward it, even when it is irrelevant.
    The same facts framed as gains or losses can flip a choice.
    Present bias makes plans and actions disagree; commitment devices exist because people know it.
    The default is a choice someone made for you; when people are inattentive, it decides.
    People will pay to punish unfairness, and fairness norms differ across societies.
    Ask of any famous finding: replicated? how big? in whom? Direction survives more often than size.
  

  Where it shows up next
  
    Consumer Behavior (34/58): reference prices, anchoring on list prices and "was/now" framing are everyday marketing.
    Pricing (34/61): decoys, round numbers and subscription defaults rely on these effects, and the replication record tells you which to trust.
    Public Economics (34/49): automatic enrolment and opt-out donation are policy tools, with the nudge debate attached.
    Strategy & Competition (34/54) and Game Theory (34/45): the ultimatum game is a game theory result that real people do not follow.
    In the world: pension design, fundraising, health messaging, app design and every "limited time offer".
  

  If you want to go deeper
  
    📘 Textbook. Nick Wilkinson and Matthias Klaes, An Introduction to Behavioral Economics (Palgrave). A standard undergraduate text.
    📘 Reference. Colin Camerer, George Loewenstein and Matthew Rabin (eds.), Advances in Behavioral Economics (Princeton, 2004), for the research papers behind the field.
    📘 Gentler. Richard Thaler, Misbehaving: The Making of Behavioral Economics (Norton, 2015), and Daniel Kahneman, Thinking, Fast and Slow (2011). Read both with the report card in mind, as some of the studies in the second have since been questioned.
    🌐 Free course. MIT OpenCourseWare 14.13 Psychology and Economics (Spring 2020, Prof. Frank Schilbach).
  

  
    The takeaway
    Real choosers rely on shortcuts, judge from reference points, discount the immediate over the later, follow defaults and care about fairness. Those departures are real and often predictable, but sizes are smaller and less universal than the first headlines claimed, and a few famous results did not survive. Trust the direction, doubt the magnitude, and check the replication.
    Try this: this week, find one default in your life (a subscription, a notification setting, a pension contribution) and decide deliberately whether it is the one you would have chosen.
  

  
    Sources & Further Reading
    Sources. Simon on bounded rationality and satisficing (1950s); Tversky & Kahneman, "Judgment under Uncertainty: Heuristics and Biases," Science 1974 (wheel-of-fortune anchoring: anchors 10 and 65, typical answers about 25% and 45%); Kahneman & Tversky, "Prospect Theory: An Analysis of Decision under Risk," Econometrica 47(2), 1979; Kahneman won the 2002 economics Nobel. Value-function parameters α ≈ 0.88 and λ ≈ 2.25 are the Tversky & Kahneman (1992) estimates; they vary by sample and are used here only as an illustration. Tversky & Kahneman, "The Framing of Decisions and the Psychology of Choice," Science 1981. Thaler, "Some Empirical Evidence on Dynamic Inconsistency," 1981 ($15 now against $30 in 3 months, $60 in 1 year, $100 in 3 years); Ainslie's $50/$100 reversals; Laibson, "Golden Eggs and Hyperbolic Discounting," 1997. Madrian & Shea, NBER Working Paper 7682 (2000), published 2001: participation "significantly higher under automatic enrollment." Organ-donation contrast and the caveat about other causes: Wikipedia, "Organ donation." Thaler & Sunstein, Nudge (2008). Nudge effectiveness: Mertens et al. (PNAS 2022) found moderate effects; Maier et al. (PNAS 2022) found no evidence after adjusting for publication bias; DellaVigna & Linos (Econometrica 2022) found smaller effects in nudge-unit trials than in published papers. Güth, Schmittberger & Schwarze, 1982; Oosterbeek, Sloof & van de Kuilen 2004 (37 studies: offers about 40%, offers below about 30% often rejected); Henrich et al. on small-scale societies. Replication: Open Science Collaboration 2015 (36%); Camerer et al. 2016 (11 of 18) and 2018 (13 of 21); ego depletion: Hagger et al. 2016 (2,141 participants, no evidence) and Vohs et al. 2021 (d = 0.06); marshmallow test: Mischel & Ebbesen 1970 (32 children), Watts, Duncan & Quan 2018; loss aversion: Gal & Rucker 2018, Mrkva et al. 2020, Yechiam & Zeif 2025; choice overload: Iyengar & Lepper 2000, Scheibehenne et al. 2010, Chernev et al. 2015. Computed in node from the code on this page: break-even win in the coin flip = λ1/α × 100 = 251 (200 at λ = 2, α = 1); anchoring index 20 ÷ 55 = 0.36; bat and ball 5p; the toy ultimatum game's best offers (39 and 3).
    
      🌐 MIT OCW 14.13 Psychology and Economics
      📘 Thaler, Misbehaving
      📘 Kahneman, Thinking, Fast and Slow


Step 7. Marketing Research & Analytics: Asking, Testing and Not Fooling Yourself
--------------------------------------------------------------------------------
Why: Asking customers, testing on them, and reading the answer correctly.
Every survey, test and dashboard gives you a number. The skill is knowing how the question steered it, how much noise is in it, and what it quietly stopped measuring.

Monday morning, the dashboard is green. The new checkout button, variant B, "converted 12% better" over the weekend. Someone wants to roll it out to everyone and put it in the quarterly slides. Before anyone does, a marketing analyst asks four quiet questions. How many people actually saw each button? Could a lift this size appear when the two buttons are identical? Who was in the test, and who was missing? And is "conversion" still the thing the company wants, or just the thing it counts?
  Those four questions are most of this subject. Marketing research is the craft of finding out what customers think, want and do, without fooling yourself on the way. It sits between statistics and psychology, and it is the course where the statistics you learned meet people who do not behave like coin flips: they misread your question, skip your survey, and change what they do once they are measured.

  
    The course in one box
    
      Who takes itMarketing majors, usually in year 2 or 3, after the first marketing course and a statistics course. Business schools also teach it as "marketing analytics" in MBA programmes, and data-science students meet it as experimentation.
      What it is forTurning a business question ("should we change the price, the pack, the page?") into evidence you can trust, and knowing how far to trust it.
      Comes afterMarketing Management (34/57) and Probability & Statistics (34/05).
      Leads toDigital Marketing & Advertising (34/62) and Product Innovation & Go-to-Market (34/63). It also feeds Econometrics (34/44), where the question "what did the campaign cause?" gets its full treatment.
      What this piece gives youYou will not be able to run a conjoint study or a clustering project professionally. You will see why a leading question moves an answer, why a bigger sample does not fix a biased one, why a test "wins" by luck so often, and why a target ruins a metric. Easier first reads: Statistics for Reading the World (12/05), How to Talk to Customers (Before You Build Anything) (04/05) and The Biases That Fool Everyone (17/02).
    
  

  Asking well: a question steers its own answer
  A survey question is not a window onto what people think. It is a nudge, and people answer the question they hear. A famous demonstration: Elizabeth Loftus and John Palmer showed people films of car crashes in 1974 and asked how fast the cars were going when they "smashed into" each other. Other groups heard "hit" or "contacted". The more violent the verb, the faster people said the cars had been going. The wording did not change the film. It changed the memory report.
  Researchers keep a short list of ways a question goes wrong. A leading question smuggles in the answer ("How satisfied are you with our excellent service?"). A double-barrelled question asks two things at once, so a "no" cannot be read ("Is the price and quality good?"). A loaded option set has more positive choices than negative ones. And a hypothetical question ("Would you buy this?") invites a cheerful yes that real wallets rarely honour. The fix is boring and reliable: ask about what people did (last purchase, last week), one thing at a time, with balanced options, in plain words, and test the draft on five people before sending it to five thousand.
  
    Interactive · Which wording is fair?
    Four pairs. Tap the wording a careful researcher would send.
    
    Score: 0 of 0 answered.
  

  Talking before counting: qualitative research
  Surveys can only measure what you already thought to ask. Before you count anything, you listen: one-to-one interviews, watching someone use the product, a focus group, a diary study. This is qualitative research. It gives you words, not percentages, and its job is to find the questions worth putting in the survey and the hypotheses worth testing. A bakery that "knows" customers want cheaper bread might hear in three interviews that the real problem is that the bread goes stale by Tuesday.
  How many people is enough? Usability researchers model it simply: if each tester independently runs into any given problem with probability p, then n testers find a share 1 − (1 − p)n of the problems. Jakob Nielsen and Thomas Landauer's 1993 paper put the average p near 0.31 for usability problems, which is where the famous "five users find about 85% of the problems" rule comes from. Treat that as a rule of thumb for obvious problems in one product, not a law: rare problems and different user types need more people, and a focus group's loudest voice can bury the rest.
  share found = 1 − (1 − p)np = chance one tester hits a given problem · n = testers
  
    Interactive · How many interviews?
    Chance one person hits a given problem (p)
      
    Number of people
      
    
    
    A model, not a measurement. 0.31 is the average Nielsen and Landauer reported; your product and your users will differ.
  

  Sampling and the margin of error
  You cannot ask everybody, so you ask a few and infer the rest. A properly random sample of n people gives an estimate of a proportion that is off by roughly this much, 19 times in 20:
  margin of error ≈ 1.96 × √( p(1−p) ÷ n )p = the share you measured, n = people who answered. Worst case p = 0.5.
  For n = 1,000 and p = 0.5 that is ±3.1 points. Quadruple the sample to 4,000 and the margin only halves, to ±1.5, because n sits under a square root. That is why national polls stop at around a thousand: after that, money buys little.
  The formula hides a trap. It assumes the sample is random: that everyone had the same chance of ending up in it and answering. When that fails, a bigger n only makes you more confident in the wrong number. In 1936 the magazine Literary Digest mailed about 10 million ballots and got about 2.38 million back, a vast sample, and predicted that Alf Landon would beat Franklin Roosevelt with about 57% of the vote. Roosevelt won with 60.8%. The mailing lists came from its own readers, car owners and telephone subscribers, who in the Depression were richer than average, and later research pointed to non-response as the main cause of the size of the miss: people angry at Roosevelt were more likely to send their ballot back.
  
    Interactive · Forty polls of the same town
    True share who like the product
      
    Respondents per poll
      
    Fans answer this many times as often as others
      
    Run 40 new pollsAdd a response bias of 1.4
    
    
    Each bar is a poll's 95% interval. The dashed line is the truth. Illustrative simulation, not real data.
  
  With no bias and 257 respondents, about 38 of the 40 bars cover the truth. Raise the sample to 3,000 with a 1.4 response bias and the bars get tight and none of them covers the truth. Precision and accuracy are different things: sample size cures noise, never bias.

  A/B tests: how often nothing "wins"
  An A/B test shows version A to a random half of visitors and version B to the other half, then asks whether the difference in conversion is more than luck. Randomising is what makes it a real experiment: any other difference between the groups is chance, so a difference in outcome can be pinned on the change. The test statistic is a z-score: the observed gap divided by the gap's typical size from noise alone. Past ±1.96, a gap this big would occur by luck about 1 time in 20 if the versions were identical (p < 0.05).
  Notice what "1 in 20" means. Run 20 tests on changes that do nothing and you expect one to "win". Now add the common mistake, peeking: you look at the dashboard every day and stop the moment it turns significant. Each look is another chance for luck to cross the line, and the false-alarm rate climbs well above 5%. Statisticians have known this since sequential analysis in the 1940s; online experimentation books still list it as the number-one error. Run an A/A test (two identical versions) below and watch.
  
    Interactive · Does the test cry wolf?
    
      Read once, at the end
      Peek 10 times, stop at first "win"
    
    Visitors per version
      
    Baseline conversion
      
    True lift of B over A (0 = identical)
      
    Run 400 experiments
    Press the button.
    Each experiment simulates real visitors with a seeded random generator, then applies a two-sided z-test at 5%. Illustrative simulation.
  
  With identical versions and one look, about 5% are flagged: that is the promise. With ten looks, roughly 18% are. And with a real lift of 10% on a 5% baseline and 2,000 visitors per version, most tests miss the real effect: small samples have low power. Good practice is to decide the sample size in advance from the smallest lift worth caring about, run to that size, and read once; or use a method designed for continuous monitoring.

  Conjoint: what people trade off
  Ask "which of these features matters to you?" and everyone says all of them, especially the cheap price. Real choices are trade-offs, so conjoint analysis forces them. You describe products as bundles of attributes (price, battery life, waterproofing), show people many bundles, and have them rate or choose among them. Because the bundles vary in a planned way, a little arithmetic pulls out how much each level contributed to liking: its part-worth. The method comes from mathematical psychology (Luce and Tukey, 1964) and reached marketing in the 1970s. Today's version mostly shows pairs or triples and asks for a pick, but the logic is the same.
  Below, a made-up person has rated eight portable speakers from 0 to 10. Two prices, two battery lives, waterproof or not: all eight combinations, which makes the design balanced, so each part-worth is simply the average rating when an attribute is at one level minus the average at the other. The ratio of a feature's effect to the price effect turns it into dollars.
  
    Interactive · A mini conjoint you take
    
      Price-minded
      Feature-lover
      Price-blind
    
    
    
    Drag the sliders to your own taste. Importance = an attribute's range of effect ÷ the sum of ranges. Dollar value = feature effect ÷ (price effect per dollar). Illustrative, eight profiles only.
  
  Try the "feature-lover": the dollar values come out far above the price range tested. That is the classic warning: conjoint works inside the range you tested and stretches badly outside it. It also measures stated choices; it does not know what happens on a crowded shelf, so serious studies check it against real sales.

  Segmentation: clustering finds the groups, you name them
  "The customer" is a fiction. A gym has early-morning regulars and January joiners; a supermarket has stock-up families and lunchtime grazers. Segmentation splits a market into groups who respond alike, so you can serve each differently. When you have data on behaviour, a standard method is k-means clustering: pick k, drop k centres at random, assign every customer to the nearest centre, move each centre to the average of its customers, and repeat until nothing moves.
  Two facts to remember. First, you choose k, and the data will always give you k groups, real or not. The honest check is the "elbow": how much tighter the groups get as k rises. Second, the result depends on where the centres start; a bad start can settle on a worse split. The algorithm finds clusters; a human has to decide whether they mean anything, and give them names like "bargain hunters" that a campaign can act on.
  
    Interactive · Step a clustering
    Number of groups (k)
      
    One stepRun to the endNew random start
    
    
    90 invented customers. Horizontal: visits per month. Vertical: spend per visit. Spread (inertia) = total squared distance to each group's centre.
  

  Dashboards and the metrics trap
  A dashboard turns a messy business into a few numbers, and then people steer by them. In 1975 the British economist Charles Goodhart noted that a statistical regularity tends to collapse once pressure is put on it for control purposes. The anthropologist Marilyn Strathern put it in the form everyone quotes in 1997: when a measure becomes a target, it ceases to be a good measure. Goodhart himself later said he meant it as a throwaway remark.
  The marketing version is everywhere. Reward a team for clicks and headlines get more tempting and less honest. Reward a bank's staff for accounts opened and, in the best-documented case, they open accounts nobody asked for: in September 2016 US regulators announced $185 million in penalties against Wells Fargo after its own analysis found employees may have opened more than two million deposit and credit card accounts without authorisation, to meet sales targets and earn bonuses. The measure, accounts per customer, had become the target.
  The defence is a habit, not a formula: pair every metric with one it can't easily fake (clicks with purchases, purchases with returns and complaints), look at the distribution as well as the average, and ask what you would do to hit this number if you did not care about the customer.
  
    Interactive · Push for clicks
    How hard the campaign chases clicks
      
    
    
    A stylised model: curiosity headlines raise clicks (1,000 × (1 + 1.5x)) but the share of clickers who buy falls (6% × (1 − 0.85x²)), with x the pressure from 0 to 1. The shape is the lesson, not the numbers.
  

  What a graduate just sees
  
    Every question steers; the fairest ones ask about past behaviour, one thing at a time.
    Interviews find the questions; surveys count the answers. The first few interviews find most of the obvious problems.
    A margin of error shrinks with the square root of n, and it only describes noise. Bias needs a different cure.
    One test in twenty on a do-nothing change will "win"; peeking makes it far worse.
    "Which features do you want?" is the wrong question; make people trade off.
    Clusters are found by a machine and named by a person; k is a choice.
    A metric that is also a reward is already being gamed.
    An unusually good result is more likely a measurement problem than a miracle.
  

  Where it shows up next
  
    Digital Marketing & Advertising (34/62): the funnel, attribution and incrementality testing are this subject at ad-platform scale; every online ad is a running experiment.
    Product Innovation & Go-to-Market (34/63): concept tests, conjoint and pre-launch surveys decide what ships.
    Econometrics (34/44): when you cannot randomise, how do you still say what caused what?
    In the world: market-research firms, product analytics teams, political polling, the "experimentation platform" inside every large app, and any role that starts with the words "insight" or "growth".
  

  If you want to go deeper
  
    📘 Textbook (analytics). Wayne Winston, Marketing Analytics: Data-Driven Techniques with Microsoft Excel (Wiley, 2014). Worked spreadsheet examples of conjoint, segmentation and forecasting.
    📘 Textbook (experiments). Ron Kohavi, Diane Tang and Ya Xu, Trustworthy Online Controlled Experiments (Cambridge University Press, 2020). Written by practitioners at Microsoft, Google and LinkedIn; best on peeking, power and metric design.
    📘 Gentler. Darrell Huff, How to Lie with Statistics (1954). Short, funny, and still the quickest tour of misleading samples and graphs.
    🌐 Free course. MIT OpenCourseWare 15.071 The Analytics Edge (Spring 2017): data-driven cases with video lectures. Khan Academy's free Statistics and Probability course covers the margin of error and significance tests.
  

  
    The takeaway
    Evidence is made, not found. The wording of the question, the way the sample was drawn, the number of times you looked, and the target you set all shape the number before you see it. A marketing analyst's real skill is asking, calmly and before the meeting, how each of those shaped this one.
    Try this: the next time you see a poll, a review score or an "up 12%" claim, find three things: how many people, how they were chosen, and what exactly was asked or counted. If you can't find them, treat the number as decoration.
  

  
    Sources & Further Reading
    Sources. Margin of error, z-tests, the A/B simulation and k-means are standard statistics, computed here and checked in code. Loftus and Palmer, "Reconstruction of automobile destruction", Journal of Verbal Learning and Verbal Behavior (1974). Nielsen and Landauer, "A mathematical model of the finding of usability problems", INTERCHI 1993. The 1936 Literary Digest figures are from the Wikipedia article on the magazine, which summarises later analyses. Luce and Tukey, "Simultaneous conjoint measurement", Journal of Mathematical Psychology (1964). Goodhart (1975) and Strathern, "'Improving ratings': audit in the British University system", European Review 5(3), 1997. The Wells Fargo figures are from the US Consumer Financial Protection Bureau's press release of 8 September 2016. All datasets in the interactives are invented for illustration.
    
      🌐 MIT OCW 15.071 The Analytics Edge
      🌐 CFPB: Wells Fargo penalty announcement, 2016
      📘 Kohavi, Tang and Xu, Trustworthy Online Controlled Experiments
      📘 Huff, How to Lie with Statistics


Step 8. Organizational Behavior & Leadership: People Are the System
-------------------------------------------------------------------
Why: Marketing is teams, agencies and persuasion inside the firm too.
Why some teams speak up and others go quiet, what actually moves people to work, and which famous findings about people at work did not survive being repeated.

Two teams get the same brief, the same budget and equally talented people. One ships, learns and enjoys it. The other argues in the corridor, nods in meetings and quietly fails. Nothing in the spreadsheet explains the difference. Anyone who has worked in both knows it is real.
  Organizational behavior is the study of that difference: what moves people to work, how groups think, why some teams speak up and others don't, who gets listened to, and what it takes to lead. It is also a field that has had a rough decade. Several of its most famous findings did not survive being re-run. This piece teaches the parts that held up, and says plainly which famous ones did not.

  
    The course in one box
    
      Who takes itBusiness students, usually in year 2, under names like Organizational Behavior, Management or Leadership. Management is part of Wharton's Business Fundamentals core, and leadership is part of the first-year Required Curriculum at Harvard Business School (see 00_MAJORS). Psychology and engineering students meet a lighter version.
      What it is forUnderstanding why people in organisations do what they do, so you can design work, run meetings and lead teams without relying on myths.
      Comes afterNothing in particular; it builds on ordinary life in groups.
      Leads toStrategy & Competition (34/54) and Entrepreneurship (34/56) in the Business Administration major; Marketing Management (34/57) lists it too, because marketing is teams and persuasion inside the firm. It links to Game Theory (34/45) for negotiation, Behavioral Economics (34/46) for biases, and Software Engineering (34/31) and Systems Engineering (34/22), where teams are the real system.
      What this piece gives youYou will not become a coach or a CEO. You will know which claims about people at work are solid, which are folklore, and how to play a negotiation. Easier first reads: The Biases That Fool Everyone (17/02), Thinking in Systems (17/06) and Negotiation for Normal People (16/10).
    
  

  Motivation: what holds up and what doesn't
  Why do people work hard? The two theories with the most evidence behind them are less glamorous than the ones on posters.
  Goal-setting theory comes from Edwin Locke and Gary Latham. Across decades of laboratory and field studies, people given a specific, difficult goal do better than people told to "do your best", provided they accept the goal, have the ability and get feedback. A 1981 review reported that about 90% of studies found the effect. It has limits: on a complex task you have never done, a specific performance target can backfire and a learning goal ("find three strategies") works better, and Ordonez and colleagues argued in 2009 that goals can narrow attention and invite cutting corners. Locke and Latham answered in print, and the debate continues.
  Self-determination theory, from Edward Deci and Richard Ryan, says that people are most motivated and well when three needs are met: autonomy (acting from choice), competence (getting better) and relatedness (belonging). Its best-known result is that paying someone for something they already enjoy can reduce their interest afterwards: Deci found it for money in 1971, while praise did not hurt. How large and general that effect is remains argued, but the practical lesson, that control crushes interest and choice sustains it, is widely used.
  What does not hold up is a surprisingly long list. Abraham Maslow's pyramid of needs is the clearest case: Maslow never drew a pyramid. It is credited to a 1960 article by Charles McDermid, and while later work supports universal human needs, the strict stair-step order is not supported. Power posing and ego depletion are the other famous casualties. Test yourself: for each classic claim, guess how it fared.
  
    Interactive · Did it replicate?
    
    
      Holds up
      Holds, with limits
      Did not hold
      Mixed or misattributed
    
    
    Next claim
    Verdicts follow the sources listed at the end. "Holds up" means supported by several independent studies or reviews, not that it is settled for ever.
  
  The habit this teaches is the main thing the course gives: when someone says "research shows", ask how big the study was, whether it was repeated by someone else, and who stands to gain from the claim.

  Decisions in groups: groupthink, and the wisdom and madness of crowds
  Groups feel wiser than individuals, and sometimes are. In 1907 Francis Galton collected about 800 guesses at a country fair in Plymouth for the weight of a slaughtered ox. The median was 1,207 pounds. The true weight was 1,198: within 1%. James Surowiecki's 2004 book The Wisdom of Crowds made the idea famous and named the conditions: diverse views, guesses made independently, and a question with a checkable answer. Averaging cancels errors that are random; it cannot cancel errors that everyone shares.
  The opposite failure is Irving Janis's groupthink (1972, revised 1982): tight, insulated groups under stress that seek harmony over truth. He applied it to the Pearl Harbor warnings, the Bay of Pigs invasion and the escalation of the Vietnam War, and listed eight symptoms such as an illusion of invulnerability and pressure on dissenters. His remedies are still good practice: a designated devil's advocate, a leader who speaks last, outside experts, and independent sub-groups. As a scientific theory it has fared less well. Lab and field evidence is mixed, and later researchers (Kramer in 1998, Baron in 2005) questioned how Janis read his case studies. Treat it as a checklist of risks, not a law.
  The crowd can also go mad when people see each other. Jan Lorenz and colleagues showed in a 2011 PNAS paper that letting people see others' estimates narrowed the spread of answers without making the average more accurate. You can reproduce the logic below: each simulated person starts with their own noisy guess of the ox's weight, then some of them move toward the running average of earlier guessers.
  
    Interactive · Guess the ox
    Crowd size
    How much people lean on earlier answers
    
    New crowd
    
    Illustrative model: each person's private guess is the true weight (1,198 lb) times a random factor with about 25% typical error, no shared bias. With influence w, a person's answer is (1 − w) × own guess + w × the average of all earlier answers. Not data from a real study.
  
  Turn the influence up: the dots huddle together and the crowd looks sure, yet its average gets worse, because early guessers set the tone for everyone. Agreement is not evidence. That is why good decision processes collect opinions before discussion.

  Teams: psychological safety
  The most useful single idea in modern team research is Amy Edmondson's psychological safety: the shared belief that you can take an interpersonal risk, such as asking a question, admitting a mistake or disagreeing with the boss, without being punished or humiliated. She defined it in a 1999 study of work teams (Administrative Science Quarterly 44(2)) after a puzzle in hospitals: in an earlier study, units with better leadership appeared to make more errors. The explanation was that people on those units felt able to report them. Teams that report nothing are not safer, only quieter.
  Google's internal Project Aristotle, begun around 2012, went looking for what made its own teams effective, expecting to find the right mix of skills and personalities. What it reported instead was that who was on the team mattered less than how members treated each other. Its write-up named five dynamics, with psychological safety the most important: then dependability, structure and clarity, meaning, and impact. Be careful about what that proves. It is one company's internal analysis, not a controlled experiment, and the safety link is correlational. Independent academic reviews do support a positive link between psychological safety and learning and performance, but "safe" does not mean "soft": the strongest teams also hold high standards.
  P(hear it) = 1 − (1 − p)kk people notice a problem, each speaks up with chance p
  Here is the simple arithmetic of a quiet team. Suppose one junior person spots a problem and would speak up only 30% of the time. The team hears about it 30% of the time: seven times in ten, the problem stays hidden. If three people notice it, each at 30%, the chance rises to about 66%. Safety is not niceness; it is the difference between a team that hears its own bad news and one that hears it from customers.
  
    Interactive · Does the team hear about the problem?
    Chance each person speaks up
    People who noticed
    Meetings before it is too late
    
    
    A toy model: independent people, same chance every meeting. Real speaking-up is not independent (one person's silence teaches the next). Illustrative.
  

  Power and influence
  People often follow because they must, can be rewarded, admire someone or simply believe the person knows. In 1959 the social psychologists John French and Bertram Raven sorted this into five bases of power: legitimate (the role gives the right to ask), reward, coercive (the power to punish), expert (knowing more) and referent (being liked or admired). Raven later added a sixth, informational power. The list is a taxonomy, not a law, but it is a good diagnostic: the base decides how long the influence lasts and what it costs. Coercion works only while you are watching; expertise and admiration work when you are not.
  
    Interactive · Name the base of power
    
    
      Legitimate
      Reward
      Coercive
      Expert
      Referent
    
    
    Next scenario
  
  Position in a network matters as much as rank: the person who connects two groups, or who everyone asks for help, has influence no title gives. And one popular shortcut did not survive: the idea that standing in a "power pose" changes your hormones and your behaviour. Co-author Dana Carney said in 2016 she no longer believed the effects were real. Looking confident may help how you feel; it does not rewrite your biology.

  Culture: what gets rewarded is what you get
  Edgar Schein, whose 1985 book on organizational culture is a standard text, described culture as three layers. On top are artifacts: dress, office layout, rituals, slogans. Beneath are espoused values: what the company says it believes. At the bottom are basic assumptions: the unspoken rules people actually follow, which you learn by watching what happens when values and targets collide. When the stated value is "safety first" and the bonus is paid on output, the assumption is output first.
  Steven Kerr's 1975 article "On the Folly of Rewarding A, While Hoping for B" is the classic version: organisations announce teamwork and pay for individual results, announce quality and pay for volume, then wonder why people do what is paid for. The practical test of a culture is the answer to a few questions. Who got promoted last year? What was the last bad news someone delivered, and what happened to them? What would a new hire learn from the first week about what is tolerated? Culture is not a poster; it is what the reward system, the leaders' behaviour and the stories people tell all make rational.

  Leadership: what the evidence says
  The oldest leadership experiment is Kurt Lewin, Ronald Lippitt and Ralph White's 1939 study of children's clubs led in autocratic, democratic and laissez-faire styles. It popularised the three styles and the finding that democratic groups were generally more satisfied. It was small, and adults were acting out styles; it is a landmark, not a proof of a best style.
  Later research moved on. Trait searches (are leaders taller, more extroverted, more intelligent?) found modest links at best. Contingency theories, such as Fred Fiedler's of the 1960s, argued that the best style depends on the situation. The model with the most research behind it is transformational leadership, from James MacGregor Burns and Bernard Bass: leaders who inspire a shared vision, consider individuals and challenge assumptions. Meta-analyses such as Judge and Piccolo's (2004) find it linked to followers' satisfaction and to leaders' rated effectiveness, though contingent reward, a "transactional" practice of clear rewards for results, also shows positive links. Critics (including Daan van Knippenberg and Sim Sitkin, 2013) point to a problem in the measures, which overlap with the outcomes they predict, and to the difficulty of showing that the style causes the result rather than following success.
  What a reader can safely carry away: leaders matter, but less than heroic stories suggest; clear direction, fairness, support and honest feedback have the most consistent backing; and people credit leaders for results they only partly controlled, a bias James Meindl and colleagues called the "romance of leadership" in 1985.

  Negotiation basics: know your BATNA
  In 1981 Roger Fisher and William Ury, of the Harvard Negotiation Project, published Getting to Yes. Its central tools are three. Separate people from the problem. Focus on interests (why you want something) rather than positions (what you say you want). And know your BATNA: your Best Alternative To a Negotiated Agreement, the best thing you can do if the talks fail. The idea echoes the "disagreement point" in John Nash's bargaining theory.
  Your BATNA is your real power, and also your walk-away line: never accept less than it. Your reservation price is the worst deal you would take; the other side has theirs. If your reservation price is below theirs, a zone of possible agreement (ZOPA) exists between the two, and the negotiation is about how to divide it. If not, no deal is available and the right move is to walk. Research on anchoring (Galinsky and Mussweiler, 2001) finds that whoever makes the first offer often pulls the final price their way, which is why preparation matters.
  
    Interactive · Sell the car
    
      Dealer offers $3,500
      Dealer offers $4,500
      Dealer offers $5,500
    
    Your asking price
    Ask this priceAccept buyer's bidTake the dealer's offerNew buyer
    
    A teaching model. The buyer has a hidden limit between $4,000 and $7,000. Each round the buyer closes 30% of the gap to that limit, and accepts any ask up to halfway between its bid and its limit. Real buyers are less predictable. Illustrative.
  
  Play it a few times. The best asks sit just under the buyer's limit but not above it, which you can only judge from how the bids move, and sometimes the buyer's limit is below your dealer's price, so the winning move is to leave. Negotiation is mostly preparation: knowing your alternative before the conversation starts.

  What a graduate just sees
  
    "Research shows" is a prompt to ask: how large, repeated by whom, and what survived?
    Specific goals help, but narrow goals can corrupt what they don't measure.
    Control kills interest; choice, competence and belonging sustain it.
    A group that agrees fast is not necessarily right; collect opinions before the discussion.
    Silence is not safety. A team that reports no problems may be one that cannot.
    Ask which base of power a person uses: it tells you how long their influence will last.
    Culture is what gets rewarded and tolerated, not what is on the wall.
    Know your BATNA before you sit down; it is your walk-away line and your leverage.
  

  Where it shows up next
  
    Strategy & Competition (34/54) and Entrepreneurship (34/56): a strategy has to be carried out by people, and a start-up is mostly a team surviving uncertainty.
    Behavioral Economics (34/46) and Game Theory (34/45): the biases and the bargaining model under the people-side lessons here.
    Software Engineering (34/31) and Systems Engineering (34/22): code reviews, post-mortems and design reviews work or fail on psychological safety.
    Marketing Management (34/57): persuasion inside the firm, agency relationships and the buyer-seller negotiation.
    In the world: hiring and promotion, running meetings, performance reviews, mergers, hospitals and airline crews, and every salary talk you will ever have.
  

  If you want to go deeper
  
    📘 Textbook. Stephen Robbins and Timothy Judge, Organizational Behavior (Pearson). The most widely assigned undergraduate text, with chapters on every topic here.
    📘 Textbook. The free OpenStax Organizational Behavior text, written for introductory courses and available online at no cost.
    📘 Gentler. Amy Edmondson, The Fearless Organization (Wiley, 2018) on psychological safety; Roger Fisher and William Ury, Getting to Yes (1981) on negotiation.
    🌐 Free course. MIT OpenCourseWare 15.311 Organizational Processes (Fall 2003): a graduate Sloan course in organisations and leadership.
  

  
    The takeaway
    People at work are moved by choice, competence and belonging more than by pyramids and poses. Groups are wise when members think independently and foolish when they copy each other. Teams learn when it is safe to speak. Culture is what gets rewarded. Lead with clarity and fairness, and walk into any negotiation knowing your alternative. Trust the findings that have been repeated, and be wary of the ones that made a good story.
    Try this: at your next meeting, write down your view before anyone speaks, and ask two others to do the same. Compare. Notice how much of the later discussion simply moved toward the first confident voice.
  

  
    Sources & Further Reading
    Sources. Goal setting: Locke & Latham, "Building a practically useful theory of goal setting and task motivation", American Psychologist (2002); Ordonez et al., "Goals Gone Wild" (2009). Self-determination theory: Deci (1971); Deci & Ryan (1985). Maslow's pyramid: attributed to McDermid, Business Horizons (1960); Tay & Diener (2011). Power posing: Carney, Cuddy & Yap (2010); Ranehill et al. (2015); Carney's 2016 statement. Ego depletion: Hagger et al. (2016); Vohs et al. (2021). Groupthink: Janis (1972, 1982); Kramer (1998); Baron (2005). Crowds: Galton (1907); Surowiecki (2004); Lorenz et al., PNAS (2011). Psychological safety: Edmondson, ASQ 44(2), 1999; Google re:Work on Project Aristotle. Bases of power: French & Raven (1959). Culture: Schein (1985); Kerr, Academy of Management Journal (1975). Leadership: Lewin, Lippitt & White (1939); Burns (1978); Bass (1985); Judge & Piccolo (2004); Meindl, Ehrlich & Dukerich (1985). Negotiation: Fisher & Ury (1981); Galinsky & Mussweiler (2001). Replication facts were checked against the Wikipedia summaries of these studies, which cite the papers. Interactive numbers were computed and checked in code; the data are illustrative.
    
      🌐 MIT OCW 15.311 Organizational Processes
      📘 Robbins & Judge, Organizational Behavior
      📘 Edmondson, The Fearless Organization
      📘 Fisher & Ury, Getting to Yes


Step 9. Brand Management: What a Brand Is, and Why It Is Worth Money
--------------------------------------------------------------------
Why: What a brand is, why it is worth money, and how it is built and lost.
A brand is not a logo. It is what comes to mind when a buyer meets a cue, and it is built from memory. Marketers still argue over whether it grows by being loved or by being easy to find.

Take the colour off a bottle of your favourite fizzy drink and you could probably still name it. The red can, the curved glass, a jingle, even a sound: small cues that your memory attaches to one name. A company that spent billions on that connection owns something that does not appear on any factory floor, and it can be worth more than the factory.
  Brand management is the discipline of building, protecting and valuing that connection. It has two schools that disagree sharply about how brands grow. A graduate knows both arguments, because both are used in real meetings.

  
    The course in one box
    
      Who takes itMarketing majors in year 3, as an elective or core course (Wharton's marketing concentration lists brand among its electives, after the consumer and research courses). MBAs take it as "Strategic Brand Management".
      What it is forDeciding what a name should mean in people's minds, how to protect it, and what it is worth.
      Comes afterConsumer Behavior (34/58), which in turn needs Marketing Management (34/57).
      Leads toNo later subject lists it as a prerequisite, but it feeds Pricing (34/61), Digital Marketing & Advertising (34/62) and Product Innovation & Go-to-Market (34/63).
      What this piece gives youYou will not be able to run a brand audit. You will know what a brand is, how its value is estimated, what each school recommends, and why two well-known rebrands failed. Easier first reads in the library: Social Media Marketing (The Default Course) and The Biases That Fool Everyone (17/02). This one goes further.
    
  

  What a brand is: associations in memory
  The older, legal definition, one the American Marketing Association has used, is a name, term, sign, symbol or design that identifies one seller's goods and sets them apart from others. That is a trademark. The modern idea, from Kevin Lane Keller's 1993 paper "Conceptualizing, Measuring, and Managing Customer-Based Brand Equity" in the Journal of Marketing, goes further: a brand lives in the buyer's head. It is the set of associations linked to a name, such as "reliable", "for kids", "expensive", "where Grandma shops". Brand knowledge, Keller argued, is awareness plus those associations, and what makes a brand valuable is when the associations are strong, favourable and unique.
  That reframes the job. A company does not make a brand; it supplies cues (name, colour, pack, price, behaviour) and customers form the memory. The memory can be reinforced or damaged by anything the buyer sees, including things the company didn't intend.

  Brand equity: what the name adds, and how it is valued
  David Aaker, in Managing Brand Equity (1991), listed five sources of value attached to a name: brand loyalty, name awareness, perceived quality, associations, and other proprietary assets such as patents and trademarks. Brand equity is the extra value a product gets from carrying the name, compared with the same product unnamed. It shows up as higher prices, higher volume, or lower marketing cost for the same sales.
  Valuing it is hard, and there are several methods. The consultancy Interbrand's method, in outline, does three things: estimate the profit the business makes beyond what its capital requires, work out what share of the buying decision is down to the brand (the "role of the brand"), and then discount those earnings at a rate that is lower for stronger brands. Interbrand's Best Global Brands 2026 list (interbrand.com, checked 1 November 2026) puts Apple first at $514.6 billion, Microsoft second at $443.7 billion and Google third at $389.3 billion. Other firms use different methods and publish different values, so treat any single number as an estimate from one method. Try the logic yourself.
  brand value ≈ economic profit × role of brand × (1 + g) ÷ (r − g)r falls as brand strength rises · g = long-run growth · a simplified illustration, not Interbrand's model
  
    Interactive · What is a name worth?
    Revenue ($ millions)
    Operating margin
    Role of the brand in the purchase
    Brand strength score
    
    Illustrative inputs. Fixed in this toy: tax 25%, capital charge 5% of revenue, long-run growth g = 2%, discount rate r = 12% − 6% × strength/100. If economic profit is negative the value shown is zero. Real methods use forecasts, more factors and judgement; Interbrand, Kantar BrandZ and Brand Finance can give very different values for the same brand.
  
  Try the two levers a brand manager controls. Raise the role of the brand and the value rises in proportion. Raise brand strength and the discount rate falls, which lifts the value more than you'd expect. A strong brand is valuable partly because its earnings are seen as safer.

  Positioning and the classic school: loyalty and difference
  The school of thought taught for decades, associated with Aaker and Keller, says: choose a clear position, make the brand different from rivals in ways customers value, and build loyalty so that customers buy again and pay more. Keller's "brand resonance" model (Strategic Brand Management) climbs from salience (do they know you?) through meaning and judgements to resonance (do they feel attached?). On this view, targeting the right segment and being distinct is how brands win.
  It leads to practice you will recognise: brand personalities, "purpose", loyalty programmes, premium positioning. It is intuitive. The next school asks whether the data agree.

  The evidence-based school: be easy to think of and easy to buy
  Byron Sharp and colleagues at the Ehrenberg-Bass Institute for Marketing Science (University of South Australia) argue, in How Brands Grow (Oxford University Press, 2010), from decades of purchase data. Their claims: brands grow mostly by adding customers (penetration), including many who buy the category rarely, not by raising the loyalty of a few; brands in a category differ far less than marketers believe; and what matters is mental availability (coming to mind in buying situations, helped by many linked cues) and physical availability (being easy to find and buy).
  Their key empirical pattern is double jeopardy, noticed by William McPhee in 1963 and developed by Andrew Ehrenberg and colleagues (for example "Double Jeopardy Revisited", Journal of Marketing, 1990): small brands have fewer buyers and those buyers are slightly less loyal. If loyalty were the engine, big brands wouldn't also have more of it by default. The surprise is that you get the pattern with no loyalty built in at all. Test it: every household picks a brand at random in proportion to its share.
  
    Interactive · Double jeopardy with no loyalty at all
    Category purchases per household
    New random households
    
    
    Invented market with five brands (shares 40, 25, 15, 12, 8%). 3,000 simulated households each buy the category the chosen number of times; every purchase picks a brand at random with probability equal to its share. No brand has loyal customers by design. Real purchases vary across households, which is why researchers use richer models.
  
  The small brand has far fewer buyers, and those buyers buy it less often per year, purely by arithmetic. Marketers who see that and say "small brands have a loyalty problem" are, on this view, misreading chance. What does the data say the small brand should do? Reach more people.
  The classic school replies. It says the evidence comes mostly from repeat-purchase packaged goods, where choices are low-involvement; that luxury, B2B, subscriptions and services may work differently; and that Sharp's sweeping rejection of segmentation and differentiation goes beyond the data. The Ehrenberg-Bass team says the patterns are found in many categories, including some services. Another widely used piece of evidence, Les Binet and Peter Field's The Long and the Short of It (2013, built on the UK IPA Effectiveness Awards database), suggests emotional brand-building drives long-term growth and activation drives short-term sales, with a rule of thumb of about 60% on brand and 40% on activation. The field has not settled the argument. A prudent reading: be easy to find and easy to remember, and be wary of any claim of loyalty without data.

  Distinctive assets: how a brand gets recognised
  If brands grow by being noticed and remembered, the cues matter: a colour, a shape, a character, a sound, a slogan. Jenni Romaniuk of the Ehrenberg-Bass Institute, in Building Distinctive Brand Assets (Oxford University Press, 2018), proposed testing each cue on two numbers: fame (what share of the category's buyers link the cue to the brand) and uniqueness (of those, what share link it only to that brand). A cue needs both. A famous but shared cue, such as red for a soft drink, helps every rival too.
  
    Interactive · Is this cue a distinctive asset?
    
      Cartoon mascot
      A common colour
      New jingle
      Generic slogan
    
    Fame: buyers who link it to the brand
    Uniqueness: of those, link it to only this brand
    
    
    Invented examples. The 50% dividing lines are chosen for this illustration; practitioners set their own benchmarks by category. The grid idea is from Romaniuk; the quadrant names here are mine.
  

  Brand extensions: borrowing a name for something new
  A brand extension uses an established name on a new product, to save the cost of building awareness. Aaker and Keller's 1990 study ("Consumer Evaluations of Brand Extensions", Journal of Marketing) found that people accepted extensions more when the parent brand was seen as high quality and the new product fit it, in three ways: it complements the parent, it can substitute for it, and the maker seems able to transfer its skills to make it. The original lab study found little sign that a bad extension damaged the parent; later research suggests dilution can happen under some conditions, so a stretch is not free of risk.
  
    Interactive · Will the name stretch?
    
      Running shoes → socks
      Soap → body wash
      Luxury watch → budget phone
      Budget airline → luxury cruise
    
    Parent brand quality (1 to 5)
    Complement fit (1 to 5)
    Substitute fit (1 to 5)
    Transfer of skills (1 to 5)
    
    Invented examples and a toy score: (quality ÷ 5) × (sum of the three fits ÷ 15). It shows the idea that quality and fit both matter, not Aaker and Keller's fitted model.
  

  Rebrands that went wrong: Tropicana 2009 and Gap 2010
  In January 2009, Tropicana's Pure Premium orange juice launched new packaging designed by the Arnell Group: the old picture of an orange with a straw in it gave way to a glass of juice and a new look. Reports from Ad Age, based on retail data, said sales of the line fell about 20% between 1 January and 22 February (about $33 million), and on 23 February 2009 the company said it would restore the old design. Several rivals gained over the same weeks. The Ehrenberg-Bass reading is that Tropicana threw away its most recognisable cue. The sales data show the fall; they cannot, on their own, separate the redesign from other causes, but the reversal speaks for itself.
  In October 2010, Gap unveiled a new logo, a bold black "Gap" in Helvetica with a small blue square, replacing its familiar white-on-navy box. After a wave of online criticism it announced a return to the old logo within about a week (sources differ on the exact days; most say the change was reversed on 11 or 12 October). Gap had also floated a crowdsourcing idea, which it dropped. No reliable figure for the cost has been published, so ignore the numbers that circulate online.
  Two lessons. First, a visual identity is a stock of memories; changing it can burn the stock. Second, rebrands often fail quietly. We remember these because the reversals were public. Not every change fails, and many go unnoticed.

  What a graduate just sees
  
    A brand is memory: associations linked to a name, not the logo.
    Brand equity is the extra value the name adds; valuing it is an estimate from a method, never a fact.
    Small brands have fewer buyers and slightly lower loyalty even in a random market: don't mistake arithmetic for a loyalty problem.
    Growth usually means reaching more people, including light buyers, not only deepening a few.
    A cue works only if it is both famous and unique to the brand.
    Extensions work when the parent is trusted and the new product fits.
    Changing an identity spends memory; do it only with evidence.
    The field is split: ask which school a claim comes from and what its data cover.
  

  Where it shows up next
  
    Pricing (34/61): a brand's premium is its price power, and discounting can erode it.
    Digital Marketing & Advertising (34/62): search and social reach (physical and mental availability online) and the long-versus-short debate.
    Product Innovation & Go-to-Market (34/63): new brand or extension? Both launches and names depend on this.
    In the world: a founder choosing a name, a company renaming after a merger, a nonprofit asking why donors recognise it.
  

  If you want to go deeper
  
    📘 Textbook. Kevin Lane Keller and Vanitha Swaminathan, Strategic Brand Management (Pearson, 5th edition). The standard course text for the classic school.
    📘 Textbook. David Aaker, Building Strong Brands (1996) and Managing Brand Equity (1991). The sources of the equity framework.
    📘 Gentler (the evidence-based school). Byron Sharp, How Brands Grow (Oxford University Press, 2010). Readable; read it alongside a classic-school book to see both sides.
    📘 Also. Les Binet and Peter Field, The Long and the Short of It (IPA, 2013). Jenni Romaniuk, Building Distinctive Brand Assets (Oxford University Press, 2018).
    🌐 Free course. MIT OpenCourseWare 15.810 Marketing Management (it includes brand sessions; I did not find a free full course on brand management).
  

  
    The takeaway
    A brand is a pattern of memories attached to a name, and it is worth money because it makes purchases easier and prices stickier. Be easy to find and easy to remember, keep the cues that people already link to you, extend only where the fit is real, and treat any brand-valuation number or growth theory as one school's view.
    Try this: cover a brand's name on a pack or ad and ask yourself what tells you who it is. Count the cues. Which are unique to it, and which could belong to any rival?
  
  
    Sources & Further Reading
    Sources. Keller, Journal of Marketing 57 (1993). Aaker, Managing Brand Equity (Free Press, 1991). Aaker & Keller, Journal of Marketing 54 (1990). Sharp, How Brands Grow (OUP, 2010). Ehrenberg, Goodhardt & Barwise, "Double Jeopardy Revisited", Journal of Marketing 54 (1990); McPhee (1963). Romaniuk (OUP, 2018). Binet & Field (IPA, 2013). Interbrand, Best Global Brands 2026 (interbrand.com, accessed 1 November 2026; the method outline is from Interbrand's description, simplified here). Tropicana: Ad Age data via Bloomberg (2 April 2009) and trade reports, 20% sales fall 1 January to 22 February 2009, reversal announced 23 February 2009. Gap: reports in The Daily Beast, Fashion Network and others from October 2010; the dates vary by one day or two between sources. The valuation, double-jeopardy and extension tools use invented data; calculations were checked in node. This is not business or investment advice.
    
      🌐 MIT OCW 15.810 Marketing Management
      🌐 Interbrand, Best Global Brands
      📘 Keller & Swaminathan, Strategic Brand Management
      📘 Sharp, How Brands Grow


Step 10. Pricing: The Number That Brings the Money In
-----------------------------------------------------
Why: The price is the one part of the mix that brings money in.
Every other choice in marketing spends money. The price is the one that collects it. Behind a good price sit a curve, a ratio, and a few tricks, some of which turn out to be less solid than the books say.

Two seats on the same flight, same row, same plane. One passenger paid $89 and the other $540. Neither is being cheated, and the airline has not made a mistake: the plane was priced by a computer that knows, minute by minute, how many people are still likely to book and how much each kind of traveller will bear. Walk through a supermarket and the same mind is at work in smaller ways: $2.99 instead of $3, a "premium" size next to the ordinary one, a coupon that only some people clip.
  A price looks like a fact about a product. It is really a decision about who will buy it, how many, and how much of their own happiness you are going to collect. The marketing and economics behind that decision are the subject here, and it is one of the few areas where a little theory is worth money the same afternoon.

  
    The course in one box
    
      Who takes itMarketing and business majors, usually in year 3 as an elective or a required "pricing" course; MBA programmes treat it as a core of marketing strategy. Economics majors meet the theory in microeconomics and the evidence in behavioural economics.
      What it is forChoosing prices, price structures and discounts that fit what customers will pay, what competitors do, and what the product costs.
      Comes afterMicroeconomics (34/42) and Consumer Behavior (34/58).
      Leads toIt is a capstone for the marketing mix and feeds Strategy & Competition (34/54), Product Innovation & Go-to-Market (34/63) and Entrepreneurship (34/56).
      What this piece gives youYou will not be able to run a pricing study. You will see why price is a choice of customers as much as of dollars, how to read elasticity, why airlines and apps charge different people differently, and which famous price-psychology tricks have solid evidence. Easier first reads: Pricing Without Fear (04/03), Where Money Actually Comes From (04/02) and The Biases That Fool Everyone (17/02).
    
  

  Willingness to pay: the curve behind every price
  Every customer carries a private number: the most they would pay for your product. It is their willingness to pay. Line customers up from the highest number to the lowest and you have drawn the demand curve: at any price, everyone whose number is at least that price buys. Raise the price and you lose the customers at the bottom of the line; lower it and you win more, but each one pays less. The seller's problem is to pick the point on that line where the money collected is largest.
  The gap between what a customer would have paid and what they did pay is their consumer surplus. Pricing is partly a negotiation over who keeps it. In the model below, 200 invented customers have willingness to pay that spreads out in the way many real markets do: a few who would pay a lot, a crowd who would pay a little. Move the price and watch revenue, profit and surplus pull against each other.
  
    Interactive · Price a product for 200 customers
    Price
      
    Unit cost
      
    How much customers value it (average willingness to pay)
      
    Cost-plus: cost + 50%Profit-maximising price
    
    
    Willingness to pay of customer i out of 200 is −mean × ln((i + 0.5) ÷ 200): an exponential spread, illustrative. A customer buys if willingness to pay ≥ price.
  
  At the defaults, a cost-plus price (cost of $20 plus half again, $30) sells to 110 people and earns $1,100 profit. The profit-maximising price is $74: it loses more than half of those customers, and earns $2,484 from the 46 who remain, in this made-up market. Now slide "how much customers value it" up to $100 and press cost-plus again. The cost-plus price is still $30 and earns $1,480; the best price has jumped to $121 and earns $6,060. That is the whole critique of cost-based pricing in one gesture.

  Three ways to set a price, and when each lies
  Businesses reach for three anchors. Cost-plus adds a margin to what it cost to make; it is simple, and it ignores the customer completely, so it leaves money on the table when the product is worth more and sells nothing when it is not. Competitive pricing follows rivals; it is natural when products are alike (petrol, flights on one route) and weakest when yours is different. Value-based pricing starts from what the product is worth to the buyer, the savings or enjoyment it delivers, and charges a share of it.
  
    MethodStarts fromWorks whenFails when
    Cost-plusyour costcosts are high and shared by all rivals; a quote must be fastcustomers would pay far more, or far less, than cost implies
    Competitiverivals' pricesproducts are near-identicalyou are cheaper to make, or better, and still match them
    Value-basedcustomer's gainthe benefit can be put in money (hours saved, losses avoided)the benefit is vague or can't be shown to the buyer
  
  A worked number. A scheduling tool saves a clinic 10 hours a month; at $25 an hour that is $250 of staff time. A value-based price might be $60 a month: the clinic keeps most of the gain and says yes quickly. Cost-plus on a tool that costs $4 per customer a month to run would say $6. The software firm that prices at $6 has made the buyer happy and itself poor. Notice also that cost still matters, as a floor: below it, each extra sale loses money.

  Price elasticity: how much volume a price change costs
  Raise the price and sales fall, but by how much? The answer is price elasticity of demand: the percentage change in quantity for a one-percent change in price. If it is more negative than −1, demand is elastic: a price rise loses more in volume than it gains per unit, so revenue falls. If it is between 0 and −1, demand is inelastic: customers shrug, and a rise brings in more money. Petrol and insulin for a diabetic are inelastic in the short run. A particular brand of cereal, with ten near-identical substitutes on the shelf, is elastic.
  ε = (% change in quantity) ÷ (% change in price)For revenue alone, the best price is where ε = −1. With a unit cost c, a profit-maximiser sets (price − c) ÷ price = 1 ÷ |ε|.
  The second rule is the one practitioners use: the more elastic the demand, the thinner the margin. Supermarkets sell milk at slim margins (shoppers notice and compare) and gift cards at fat ones. The calculator uses a simple straight-line demand: 1,000 units at a price of zero, falling by 10 for every dollar.
  
    Interactive · Revenue-maximising calculator
    Price
      
    Unit cost
      
    
    
    Demand: Q = 1,000 − 10 × P. Elasticity at a point = −10 × P ÷ Q. Revenue peaks at P = $50; profit peaks at (100 + cost) ÷ 2. Illustrative.
  
  Set the cost to $20. At $30 you sell 700 units for $21,000 of revenue and $7,000 of profit; ε is about −0.43, so demand is inelastic and a rise helps. At $50 revenue peaks at $25,000 (ε = −1) but profit is $15,000. At $60 revenue slips to $24,000 while profit rises to $16,000 (ε = −1.5), the best. Profit-maximising prices sit above the revenue-maximising one whenever costs are positive, because every unit you don't sell also saves its cost.

  Price discrimination and versioning: one product, different customers
  A single price is a compromise: set it high and you lose the thrifty, set it low and you give a discount to people who would have paid more. Price discrimination means charging different customers different prices for essentially the same thing, to collect more of each one's willingness to pay. It comes in three flavours: by person (student, senior, local-resident prices), by time or place (peak and off-peak, airport versus high street) and by version, where you offer a menu and let customers sort themselves. Economists tell the story, often credited to the nineteenth-century French engineer Jules Dupuit, of railways deliberately leaving third-class carriages uncomfortable, not out of cruelty or thrift but to keep well-off passengers from trading down. Whatever the history, the logic stands.
  The menu version is the one you see every day: good, better, best. The trick is that the top version must be worth more to the high-value customer than the saving from the cheap one, or they will trade down. Economists call that the incentive constraint. The calculator below has two groups: 60 hobbyists who would pay $40 for a basic version or $50 for the pro, and 40 professionals who would pay $60 for basic or $120 for pro.
  
    Interactive · Build a menu
    
      One product at $50
      Basic $40, pro $100
      Pro too cheap?
    
    Basic price
      
    Pro price
      
    
    Each customer picks whichever option leaves them the most surplus, or nothing. Marginal cost zero (software). "One product" sells only the pro version. Illustrative groups.
  
  The best single price (pro only, at $50) collects $5,000. With basic at $40 and pro at $100, the hobbyists buy basic, the professionals buy pro, and revenue is $6,400. Drag the pro price to $75: the professionals still buy pro, but you are charging them $25 less than they would pay, and revenue drops to $5,400. Drag it above $100 and they trade down to basic. Versioning works only where the customers truly differ in how much they value the extra.
  Is it fair? Price discrimination is legal in most ordinary retail settings, and it can serve more people overall (the hobbyists get a version they can afford). It is restricted when it targets protected characteristics or in regulated markets such as utilities, and customers who discover they paid more for no visible reason tend to feel cheated. The honest firms hang the difference on something real: timing, quantity, features, or a cost.

  Bundling: selling two things as one
  Why does a streaming service sell hundreds of shows for one price, and a phone company a bundle of calls, data and texts? Because customers disagree about which parts they like. Take a software firm with two programs and two customers whose tastes are opposite: Ann values the writing program at $100 and the spreadsheet at $40; Ben values them at $40 and $100. Sold separately, each program's best price is $100 (one buyer) or $40 (two buyers: $80), so $100 per program and $200 in total. Sold as a bundle, both value it at $140, and one price of $140 collects $280. The bundle works because the high and low valuations cancel out and the customers look alike, which makes one price fit. Economists first described this in the 1960s and 1970s, with the Stigler (1963) film-licensing example and Adams and Yellen (1976).
  
    Customers (writer, sheets)Separate, bestBundle, best
    Ann ($100, $40), Ben ($40, $100): opposite tastes$200$280
    Ann ($100, $100), Ben ($40, $40): same tastes$200$200
  
  The second row is the warning. If the people who love one part also love the other, the bundle adds nothing. Bundling pays when valuations are negatively correlated or at least not strongly positive, and when the extra cost of giving each customer everything is near zero, as with digital goods.

  The psychology of price: what held up and what didn't
  Textbooks teach three price tricks with confidence. It is fair to ask how well each has survived testing, which is a theme of this whole series.
  
    IdeaOriginal claimWhat the record shows
    Charm prices ($9 endings)$39 sells better than $40, even though 1 cent is 1 centStrongest of the three. Anderson and Simester (2003) ran three field experiments with a retailer and found 9-endings raised demand each time, more for new items and less when "Sale" cues were used. A large literature agrees on the direction; size varies with the product.
    AnchorsAn irrelevant number shifts what you'll pay: Tversky and Kahneman (1974); "coherent arbitrariness" (Ariely, Loewenstein and Prelec, 2003)Anchoring in estimates is robust. The pricing version, that a random Social Security number shifts bids for wine and chocolate, was much weaker in later tests: Fudenberg, Levine and Maniadis (2012) found very weak effects for common goods and none for lotteries; other teams found small or null effects.
    DecoysA worse third option makes the target look better (Huber, Payne and Puto, 1982); popularised by a "web / print / print+web" magazine exampleThe most fragile. Frederick, Lee and Baskin (2014) reported that the effect largely vanishes with realistic products and descriptions. The famous magazine example comes from a classroom demonstration in a popular book, not a field test on the magazine's real customers.
  
  The lesson is not "psychology is fake". It is that the effects live in particular conditions and shrink in the wild, and that a price test on your own customers beats a story from a book. Test, and publish your own "lift" with its margin of error (see Marketing Research & Analytics, 34/59).
  
    Interactive · A decoy, if it works
    Illustrative split of 100 shoppers, not data. Add a medium size that is poor value next to the large, and suppose it nudges some people from small up to large.
    
      Small $3 · Large $5
      Small $3 · Medium $4.50 · Large $5
    
    How strongly the decoy works (people moved to Large)
      
    
  
  Notice that the decoy only pays if the effect is real and big enough. At zero effect, the decoy just adds a pointless line to the menu. It is the numbers on the slider that would need testing.

  Dynamic pricing and revenue management: selling the same seat many times
  An airline seat is perishable: at departure its value drops to zero. The airline has a fixed number of seats, a few business travellers who book late and pay a lot, and many holidaymakers who book early and pay little. If it fills the plane with cheap early bookings, the late business traveller finds no seat; if it holds seats back and they go empty, that is waste. The craft of deciding how many seats to protect for the high fare is revenue management, and it began in earnest when US airlines were deregulated in 1978 and low-cost rivals such as People Express arrived. In 1985 American Airlines responded with restricted "Ultimate Super Saver" fares, advance-purchase and non-refundable, sold only where it had empty seats; the case taught in business schools reports revenue up 14.5% and profit up 47.8% in the following year.
  protect seats until P(high-fare demand > y) = low fare ÷ high fareLittlewood's rule (1972): keep selling seats at the low fare until the next seat is worth more to a late high-fare buyer.
  That rule is the heart of it. If the high fare is three times the low fare, protect seats until there's only a one-in-three chance of selling the next one at the high price. The calculator applies it to a 100-seat plane; with the starting settings the best answer is to protect 34 seats. Real systems handle dozens of fare classes, cancellations, connections and competitors, but they rest on this idea. Hotels, rental cars, concert tickets, ride apps with surge pricing and electricity time-of-use tariffs run on the same logic, with the same tension: the firm earns more, and customers feel the unpredictability.
  
    Interactive · Protect seats for the late booker
    Seats protected for the high fare
      
    Expected late high-fare demand (average)
      
    Low fare (high fare fixed at $300)
      
    
    
    100 seats; low-fare demand is large and books first; late high-fare demand is Normal with standard deviation 10. Revenue = low × (100 − protected) + 300 × E[min(demand, protected)]. Illustrative.
  

  Subscriptions and freemium: pricing the long run
  A subscription changes the question from "what will they pay today?" to "how long will they stay?". The value of a customer is roughly what they pay each month, times your margin, divided by the monthly churn (the share who leave): a customer who pays $10 at an 80% margin and leaves at 5% a month is worth about $8 ÷ 0.05 = $160 over their life. You can spend up to that much to win them, and a half-point of churn matters as much as a price rise.
  Freemium gives a basic version away free and charges for more. It works when the free users cost almost nothing to serve, bring in others, or turn into payers at a rate that covers the rest. It fails when the free tier is generous enough that few ever need to pay, or expensive enough that they cost more than they bring. The calculator holds 1,000 free sign-ups and asks the plain question: does the paying minority cover everyone?
  
    Interactive · Does the paying minority cover the crowd?
    Paid price per month
      
    Share of sign-ups who pay
      
    Cost to serve one free user per month
      
    Monthly churn of payers
      
    
    Paid gross margin 80%. Value of a payer = price × 80% ÷ churn. Illustrative.
  

  What a graduate just sees
  
    Behind any price is a demand curve: every price is a choice of which customers to keep.
    Cost sets the floor; the customer's value sets the ceiling; the price lives between.
    Elastic demand means thin margins and a revenue fall from a rise; inelastic means the opposite.
    One price is a compromise; menus, times and bundles let different customers pay in different ways.
    A bundle wins when tastes differ and loses when everyone likes the same things.
    Charm prices have good evidence; anchors and decoys are weaker than the books say. Test on your own customers.
    Perishable inventory (seats, rooms, hours) is a protect-or-sell problem, and the price moves with the odds.
    For a subscription, churn and lifetime value matter more than the monthly figure.
  

  Where it shows up next
  
    Strategy & Competition (34/54): when rivals can copy your price instantly, price becomes a game; see price wars and tacit collusion.
    Product Innovation & Go-to-Market (34/63): launch prices (skim or penetrate) decide who your first customers are.
    Entrepreneurship (34/56): most new ventures under-price; the demand curve is the first thing to test.
    In the world: airline and hotel revenue teams, SaaS pricing pages, supermarket category managers, ride apps, and utilities regulators.
  

  If you want to go deeper
  
    📘 Textbook. Thomas Nagle and Georg Müller, The Strategy and Tactics of Pricing (Routledge). The standard practitioner text, strong on value-based pricing.
    📘 Textbook. Hermann Simon and Martin Fassnacht, Price Management (Springer, 2019). More quantitative, with European and B2B cases.
    📘 Gentler. William Poundstone, Priceless: The Myth of Fair Value (2010). A readable tour of price psychology; read it alongside the replication notes above.
    🌐 Free course. MIT OpenCourseWare 15.071 The Analytics Edge includes optimisation of revenue under demand uncertainty; CORE Econ's free textbook The Economy (core-econ.org) covers demand, elasticity and market power.
  

  
    The takeaway
    A price is a choice of who you serve and how much of their gain you collect. Cost sets the floor, elasticity says how much volume each dollar costs, and versions, bundles and clocks let you charge different people differently. The psychology of price is real but smaller and shakier than the pop books say.
    Try this: this week, spot one price menu (a coffee-cup sizes board, a streaming plan page) and work out which customer each option is built for, and which option is there only to make another look good.
  

  
    Sources & Further Reading
    Sources. Elasticity, the Lerner markup rule, Littlewood's rule (1972), bundling arithmetic and the LTV formula are standard microeconomics and revenue-management theory; every number in the interactives is computed in code from the stated assumptions and is illustrative, not data. Anderson and Simester, "Effects of $9 price endings on retail sales: evidence from field experiments", Quantitative Marketing and Economics 1(1), 2003. Fudenberg, Levine and Maniadis, "On the robustness of anchoring effects in WTP and WTA experiments", American Economic Journal: Microeconomics (2012). Frederick, Lee and Baskin, "The limits of attraction", Journal of Marketing Research 51(4), 2014. Huber, Payne and Puto, Journal of Consumer Research (1982). Tversky and Kahneman, "Judgment under uncertainty", Science (1974). Stigler, "United States v. Loew's Inc.: a note on block-booking" (1963); Adams and Yellen, "Commodity bundling and the burden of monopoly", QJE (1976). The American Airlines 1985 fare figures are as reported in the Harvard Business School case summarised on Wikipedia's Revenue management article; the Dupuit railway story is widely told and is given here as a story, not a verified quotation.
    
      🌐 MIT OCW 15.071 The Analytics Edge
      🌐 CORE Econ, The Economy (free)
      📘 Nagle and Müller, The Strategy and Tactics of Pricing
      📘 Poundstone, Priceless


Step 11. Digital Marketing & Advertising: Where Attention Is Bought Now
-----------------------------------------------------------------------
Why: Search, social, ads and attribution: where attention is bought now.
Every search, scroll and page load can trigger an auction, and every click leaves a trail. The hard part is no longer reaching people. It is knowing which of the clicks you paid for actually changed anything.

You type a few words into a search box. Before the page finishes loading, a handful of advertisers have already bid for the right to sit above the results, a ranking has been computed, and the winner's price has been fixed. All of it takes less time than a blink. Later you open a news site, and while its text arrives, an exchange somewhere offers your visit, with a profile attached, to a room of bidders. By evening, an email, a post and an ad for the same shoes have each claimed that you bought them.
  Digital marketing is the machinery behind those moments: how attention is bought, how it is earned for free, how the results are counted, and how the rules about your data keep changing. This piece takes the free-to-start ideas from the library's Social Media Marketing course and goes further, into the money and the evidence.

  
    The course in one box
    
      Who takes itMarketing majors in year 3, after the first marketing course and the research and analytics course. Business schools teach it as digital or performance marketing; most university programmes also pair it with free industry certifications from the ad platforms.
      What it is forBuying and earning attention online, and measuring what it was worth.
      Comes afterMarketing Management (34/57) and Marketing Research & Analytics (34/59).
      Leads toProduct Innovation & Go-to-Market (34/63), where these channels become a launch plan; Strategy & Competition (34/54) for the platforms' power.
      What this piece gives youYou won't be able to run a campaign account. You will see how an ad auction chooses its winner and price, why "last click" credit is wrong, what a lift test is, and how the privacy rules shape all of it. Easier first reads: Social Media Marketing, From Zero to Running Your Own Strategy (The Default Course 01), The Internet: From ARPANET to the Network in Your Pocket (29/04) and Where Money Actually Comes From (04/02).
    
  

  How ads are bought: auctions in a blink
  Almost all online advertising is sold by auction, because the thing for sale (this person, on this page, right now) exists once and has no list price. Auctions have a famous design problem: if you pay what you bid, you should bid less than the thing is worth to you, and everyone ends up second-guessing everyone. In 1961 William Vickrey described the fix. In a second-price auction the highest bidder wins but pays only the second-highest bid. Now there is no point shading: bidding exactly what it is worth to you is the best you can do, whatever the others bid.
  Search ads use a variant. Several ads share the page, ranked by Ad Rank, roughly bid times a quality score (how likely people are to click, how relevant the ad and the landing page are). You pay only when someone clicks, and the textbook rule is that you pay the least you could have bid and still held your position:
  price per click = (Ad Rank of the ad below ÷ your quality) + 1 centReal systems add more signals and rules. This is the model taught by Varian (2007) and Edelman, Ostrovsky and Schwarz (2007).
  This "generalised second price" rule is not quite Vickrey's: it no longer makes honesty the strictly best bid, but it stays stable and easy to explain, and a high-quality ad can beat a higher bidder. That is why a relevant ad can pay less than a careless one for the same spot. Try changing the bid and quality of one advertiser below.
  
    Interactive · A search auction
    You are advertiser B. Others: A (bid $4.00, quality 7), C (bid $3.50, quality 5), D (bid $5.00, quality 3). Reserve price: $0.50.
    Your bid
      
    Your quality score
      
    
    
  
  Display and video ads, the banners and clips around other people's content, use programmatic buying: when a page loads, the site's ad server offers the slot to many bidders in a fraction of a second, and the highest bid wins. For years these used second-price rules. Then, in a change Google announced on 6 March 2019 and rolled out over the following months, Google's Ad Manager moved its display and video marketplace to a first-price auction, where the winner pays what it bid, and dropped the "last look" advantage that Google's own exchange had enjoyed. Buyers who now have to guess how low they can go use "bid shading", which brings the price down to what the winner needed.
  
    Interactive · Pay-your-bid versus second price
    
      Second price
      First price
    
    What the impression is worth to you
      
    Rival bidders
      
    Your bid
      
    
    
    Model: rivals' values are equally likely anywhere from $0 to $10; they bid their value in a second-price auction and (n−1)/n of it in a first-price one (the standard equilibrium). Illustrative and idealised.
  
  The pattern to carry away: in second price, your best bid equals your value, so bidding is easy. In first price, your best bid is your value times k/(k+1), where k is the number of rivals: the more rivals, the closer to your full value. Moving to first price made auctions simpler to explain to a publisher, harder for buyers to play, and a bit more of a guessing game.

  Earning attention: search optimisation, social, content and email
  Paid ads are only the visible part. Search engine optimisation (SEO) is the work of earning a place in the unpaid results. In plain words, a search engine crawls the web, builds an index of what it finds, and then ranks pages for a query by how relevant and trustworthy they look. The original idea behind Google's ranking, described by Sergey Brin and Larry Page in 1998, was that a link from another page is a vote, and votes from pages that are themselves well-linked count for more. The engines now use many more signals and change them often, so the durable advice is dull: make pages that answer the question people actually typed, load fast, work on a phone, can be crawled, and earn links from sites people trust. Tricks that aim at the ranking system rather than the reader are a race you rent, not own.
  Social media marketing is attention earned through posting, community and creators, and rented through paid placements. The rules are set by each platform's feed, so a strategy built on one feed is a strategy built on someone else's decisions. The library's course covers the how-to; the lesson here is strategic. Content marketing (useful articles, videos, tools) builds an asset that keeps bringing people in; email is the channel you own, a list you can reach without paying a platform each time.
  Email carries a measurement lesson too. Since 2021, Apple's Mail Privacy Protection loads message images in the background for users who turn it on, so "opens" stopped meaning that a person read it. Clicks, replies and purchases are the numbers worth reading. And lists must be built with permission: the EU's rules and many countries' laws require opt-in, while US law sets different baseline rules for commercial email, so check the law where your readers live.

  The funnel: where the people go
  The funnel is the oldest picture in marketing: many people notice, fewer click, fewer still sign up, fewer buy. Each step has a conversion rate, and the chain multiplies. That multiplication hides a point most people miss. A ten-percent improvement at any stage gives the same ten percent more buyers. But the stages cost different things. An ad that earns more clicks brings more visitors and a bigger bill; a landing page that converts more of the same visitors costs nothing extra. The metric to watch is cost per acquisition (CPA): spend divided by buyers, compared with what a buyer is worth.
  
    Interactive · Build a funnel and find the lever
    Click-through rate (clicks per 100 views)
      
    Cost per click
      
    Visitors who sign up
      
    Sign-ups who buy
      
    Profit per buyer, before ad cost
      
    
    100,000 ad views. Pay per click. Illustrative numbers.
  
  At the defaults, 100,000 views give 2,000 clicks costing $1,600, 100 sign-ups and 20 buyers: a CPA of $80 against $90 of profit per buyer, a $200 gain. Raise the click-through rate by 10% and you get 22 buyers, but spend $1,760, so CPA is still $80. Raise the sign-up rate by 10% and you get 22 buyers for the same $1,600: CPA falls to $72.73. Fixing the page beats winning more clicks, until the page is as good as it will get.

  Attribution: why "last click" lies
  A customer sees a creator's video, then a banner, then an email, then searches your brand name and buys. Which touch gets the credit? The default in many tools is last click: the final touch gets everything. It is easy to compute and systematically unfair. The search at the end happens because the earlier touches created the interest; paid brand search, as we shall see, often just collects people who were coming anyway. Other rules share the credit differently: first click rewards whatever introduced the customer; linear splits equally; time decay gives more to touches near the sale; position-based gives the first and last the most. Each is a rule of thumb. None is a measurement of cause.
  
    Interactive · Five customers, five rules
    
      Last click
      First click
      Linear
      Time decay
      Position (40/20/40)
    
    CustomerTouches, in order
    
    
    Five invented journeys that each ended in a sale. Time decay halves the weight for each step back from the sale. The "experiment" row is a made-up result of the kind a holdout test could produce; it only shows how far rules of thumb can sit from it.
  

  Incrementality: the question attribution cannot answer
  Attribution tells you who was present before a sale. Incrementality asks what the ad added: how many of those sales would have happened anyway? The only reliable way to find out is an experiment: show the ad to a random group, hold it back from a matched control, and compare outcomes. That is the same logic as the A/B test in Marketing Research & Analytics (34/59), applied to advertising.
  The best-known example is eBay's. A team of economists (Thomas Blake, Chris Nosko and Steven Tadelis) worked with eBay on field experiments, published in Econometrica in 2015. In one, eBay stopped paid ads on queries for its own brand for a random subset of users; the free link still appeared. They found that ads on brand keywords had no measurable short-term benefit, because people who lost the paid link almost always clicked the free one. For non-brand keywords the story was more nuanced: new and infrequent users responded to ads, but most of the spend went to frequent users whose purchases the ads did not seem to influence, so average returns were negative. Their conclusion: the returns to paid search were a fraction of what non-experimental estimates suggested, partly because people who search for something are exactly the people likely to buy it.
  Experiments are not free. Randall Lewis and Justin Rao (2015, Quarterly Journal of Economics) examined 25 large experiments run with major US retailers and brokerages, most reaching millions of customers, and found that measuring advertising's return precisely is hard even then: ad effects are small compared with the natural swings in people's purchases, so the uncertainty around the answer stays wide. A separate study at Facebook (Gordon and colleagues, 2019) compared large randomised experiments with the observational methods advertisers commonly use, and found that the observational methods often did not reproduce the experimental answer. Many ad platforms now sell "lift studies" and geographic holdout tests for this reason.
  
    Interactive · Did the ad do it?
    Clickers who would have bought anyway
      
    Ad clicks per month
      
    
    
    Control group's buying rate
      
    Real relative lift you hope to detect
      
    
    Part 1: $0.50 per click, 5% of clickers buy, $50 of revenue per sale. Part 2: rule of thumb n ≈ 16 p(1−p) ÷ (change in rate)² for 80% power at 5% significance. Illustrative.
  
  At the defaults, the dashboard credits the ad with 200 sales and $10,000 of revenue against $2,000 of spend: a return of 5 to 1. If 90% of those clickers would have bought through the free link anyway, the ad caused only 20 sales and $1,000 of revenue, a return of 0.5 to 1. The same ad, the same click data, a tenfold difference. The second half shows the price of finding that out: detecting a 5% lift on a 2% buying rate needs roughly 300,000 people per group.

  Privacy rules, cookies and a measured future
  Almost everything above runs on identifiers: cookies, device IDs, email addresses. Governments have been regulating them for a decade, and the rules differ by country. A fair summary, dated as of the sources checked for this piece:
  
    WhereRuleWhat it does for advertisers
    EU / EEAGDPR, applying since 25 May 2018; the separate ePrivacy rules on cookiesNeeds a legal basis for processing personal data; consent must be specific and unambiguous (no pre-ticked boxes); applies to outside firms that target people in the EU; fines up to €20 million or 4% of worldwide annual turnover, whichever is higher.
    California, USCCPA (passed 2018, in force 1 January 2020), later amended by the CPRA; similar state laws in Virginia and Colorado (2021)Right to know, delete and opt out of the "sale" or sharing of data. The US has relied on state laws rather than a single national statute.
    ChinaPersonal Information Protection Law, enacted 2021Consent and purpose rules modelled in part on the GDPR, with its own state-oversight provisions.
    Brazil, IndiaLGPD (law of 2018, in force 2020); India's Digital Personal Data Protection Act (2023)Consent-based regimes for personal data, with national regulators.
    ElsewhereJapan, South Korea, Singapore, Thailand and South Africa have GDPR-influenced laws; Turkey adopted one in 2016A growing patchwork, which is why global advertisers build for the strictest region.
  
  Platforms added their own rules. Apple's App Tracking Transparency, introduced with iOS 14.5 in 2021, made apps ask permission before tracking people across other companies' apps, which reduced the cross-app data that tracking-based ads relied on. Safari and Firefox block third-party cookies by default.
  What about Chrome? This is the part that most often goes stale, so read the date. Google announced in 2020 that it would phase out third-party cookies in Chrome and replace them with its Privacy Sandbox tools, then postponed it several times. In 2025 it said Chrome would maintain its current approach of offering users a third-party cookie choice, rather than removing the cookies for everyone. On 17 October 2025, Google's Privacy Sandbox team announced it was retiring most of the Privacy Sandbox technologies, including the Topics and Protected Audience interfaces and the Attribution Reporting API, citing low adoption, while continuing work on a few such as CHIPS, FedCM and Private State Tokens. As of that last official statement I could verify (October 2025), third-party cookies were still working in Chrome by default. I could not check for anything newer. If you are reading this later, check Google's Privacy Sandbox pages and your browser before relying on it.
  The practical result for a marketer: depend less on tracking individuals, more on first-party data you collected with consent, on contextual placement (the page, not the person), and on experiments that measure lift without following anyone around.

  What a graduate just sees
  
    Ad space is sold by auction; the rule of the auction decides how bidders behave.
    In a search auction, a better ad can win a higher slot at a lower price.
    Conversion rates multiply, so fix the leakiest stage; paying for more clicks and fixing the page are different spends.
    "Last click" tells you who was there at the end, not who caused the sale.
    Clicks that came from people who were already coming are not worth buying.
    Only a holdout experiment measures incrementality, and it needs a big sample.
    An "open rate" or a "view" is a proxy; read the metric nearest the money.
    Privacy rules are set by region and by platform, and both change; build on first-party, consented data.
  

  Where it shows up next
  
    Product Innovation & Go-to-Market (34/63): launch plans choose channels using funnel maths and test budgets.
    Econometrics (34/44): when an experiment is impossible, the techniques that estimate cause from messy data.
    Computer Networks (34/27) and Databases (34/28): the plumbing that serves an ad in a blink and keeps its logs.
    In the world: performance-marketing teams, agencies, publishers' ad operations, privacy officers, and the data scientists who build bidding and measurement systems.
  

  If you want to go deeper
  
    📘 Textbook. Dave Chaffey and Fiona Ellis-Chadwick, Digital Marketing (Pearson). A standard university text covering channels, strategy and measurement.
    📘 Textbook (measurement). Ron Kohavi, Diane Tang and Ya Xu, Trustworthy Online Controlled Experiments (Cambridge University Press, 2020). Experiments at web scale.
    📘 Gentler (and sceptical). Tim Hwang, Subprime Attention Crisis (2020). A readable argument that the online ad market is shakier than it looks.
    🌐 Free courses. Google Skillshop (Google Ads and Analytics training) and HubSpot Academy (free marketing courses). For the evidence side, the eBay paper below is free to read.
  

  
    The takeaway
    Online, attention is bought by auction, judged by a funnel and credited by rules that are convenient rather than true. The honest question is always "what would have happened without the ad?", and only an experiment answers it. The rules about data, set by governments and platforms, keep moving, so the safest asset is a customer who gave you permission.
    Try this: next time an ad follows you from site to site, ask what it is claiming credit for. Would you have bought anyway? If you can answer "yes", you have just understood incrementality.
  

  
    Sources & Further Reading
    Sources. Vickrey, "Counterspeculation, auctions, and competitive sealed tenders", Journal of Finance (1961). Varian, "Position auctions", International Journal of Industrial Organization (2007); Edelman, Ostrovsky and Schwarz, "Internet advertising and the generalized second-price auction", American Economic Review (2007). Google's switch to first-price auctions: announced 6 March 2019 (9to5Google, AdExchanger and Google's Ad Manager blog); described by Marketing Dive. Brin and Page, "The anatomy of a large-scale hypertextual Web search engine" (1998). Blake, Nosko and Tadelis, "Consumer heterogeneity and paid search effectiveness: a large-scale field experiment", Econometrica (2015), working version NBER w20171. Lewis and Rao, "The unfavorable economics of measuring the returns to advertising", QJE 130(4), 2015. Gordon, Zettelmeyer, Bhargava and Chapsky, "A comparison of approaches to advertising measurement: evidence from big field experiments at Facebook", Marketing Science (2019). GDPR dates, scope, consent rules and fines: the GDPR text (Articles 3 and 83) and Wikipedia's GDPR article for other countries' laws. Chrome: Google Privacy Sandbox blog, "Update on plans for Privacy Sandbox technologies", 17 October 2025. The auction, funnel, attribution and incrementality numbers in the interactives are invented and computed in code.
    
      🌐 NBER w20171: Consumer Heterogeneity and Paid Search Effectiveness (eBay)
      🌐 Google Privacy Sandbox: update, 17 October 2025
      🌐 GDPR Article 83: fines
      📘 Kohavi, Tang and Xu, Trustworthy Online Controlled Experiments


Step 12. Strategy & Competition: Why Some Firms Keep Their Profits
------------------------------------------------------------------
Why: Marketing strategy is competitive strategy seen from the customer's side.
Most good ideas get copied until the profit is gone. Strategy is the study of what stops that, and competition law is the study of when it has stopped too well.

In Damodaran's January 2026 dataset of listed companies, the whole market kept about 9.7 cents of net profit from each dollar of sales. Software firms kept about 25 cents. Tobacco firms kept about 27. Airlines kept about 2.5 and food retailers about 1.3. Airlines carry passengers across oceans and fly them safely; grocers feed whole cities. They are no less useful than software. They are just in a game where profit is competed away.
  Why? One company makes cola and another makes steel, and both work hard. Strategy asks what lets a firm keep earning above the average, and competition policy asks what society should do when a firm keeps too much.

  
    The course in one box
    
      Who takes itBusiness students (a year-3 core course in MBA and undergraduate business programmes, usually after finance and marketing have been met); economics students as industrial organization; and marketing students, who meet it as marketing strategy. Business schools teach it through cases and frameworks; economics departments teach it through game theory and models. They are the same subject seen from two sides.
      What it is forSeeing why some industries and some firms earn well for decades, and when that profit comes from a better product and when it comes from blocking rivals.
      Comes afterMicroeconomics (34/42) and Game Theory (34/45).
      Leads toIt is the capstone of Business Administration and a late core course in Economics and Marketing. It feeds Pricing (34/61) and Entrepreneurship (34/56).
      What this piece gives youYou will not be able to write a strategy for a firm. You will have a checklist of questions that experienced strategists ask, a feel for why scale and networks create winners, and the economist's view of when a market has too few competitors. Easier first reads: Where Money Actually Comes From (04/02), The Mental Models Top Operators Use (04/07) and Capitalism (31/01). Not legal or investment advice.
    
  

  Why some industries are more profitable: Porter's five forces
  In a 1979 Harvard Business Review article, "How Competitive Forces Shape Strategy", Michael Porter made a plain claim: profit depends less on how clever a firm is than on the structure of the industry it is in. He named five forces that squeeze or protect profit:
  
    Threat of new entrants. If anyone can start, profits are copied away.
    Threat of substitutes. Something different that does the same job caps your price.
    Power of buyers. Big, informed buyers who can switch bargain your price down.
    Power of suppliers. If your inputs come from one source, they take your margin.
    Rivalry. Many equal rivals with high fixed costs fight on price.
  
  A weak force is good for the firm and a strong one is bad. A graduate looks at an industry and scores the five before saying anything about the managers.
  
    Interactive · Score an industry
    Rate each force from 1 (weak, good for profit) to 5 (strong, bad). The presets use my own illustrative ratings for five industries; argue with them. The margin shown is Damodaran's net margin for the matching industry, January 2026.
    
    New entrants
    Substitutes
    Buyers' power
    Suppliers' power
    Rivalry
    
    
  
  The model has critics, and they are worth knowing. Kevin Coyne and Somu Subramaniam argue that it assumes buyers, suppliers and rivals never collude or interact, that value comes only from structure, and that uncertainty is low. Adam Brandenburger and Barry Nalebuff added complementors (the makers of things that make your product more valuable) as a sixth force. Others argue that the model ignores what each firm itself is good at, which the resource-based view puts first. Porter himself updated the framework in a 2008 article. All sides agree that both the industry and the firm matter.

  Competitive advantage: cost, differentiation and the value chain
  Inside an industry, a firm that earns more than its rivals has a competitive advantage. Porter's Competitive Strategy (1980) said there are two basic kinds. Cost advantage: you make the same thing for less, and can either undercut rivals or keep the difference. Differentiation: you make something buyers will pay more for. He also warned of being "stuck in the middle", with neither; economists and managers still argue about how often that is true.
  A neat way to see the two together is to ask what a rival's cost lets you charge. If your product is as good as the rival's, a price war drags the price down to the rival's cost, and your margin is your cost gap. If buyers value yours more, you can charge that premium on top.
  your margin = (rival's cost + your quality premium) − your costcost gap plus premium; a negative number means the rival wins the sale
  
    Interactive · Cost, premium or both
    Two firms sell to the same buyers. A price war pushes the loser's price to its own cost. Buyers value your product $premium more than the rival's. All figures invented.
    Your cost per unit
    Rival's cost per unit
    Premium buyers pay for yours
    
    
  
  Porter's value chain (Competitive Advantage, 1985) tells you where to look: inbound logistics, operations, outbound logistics, marketing and sales, service, supported by technology, procurement, people and infrastructure. An advantage hides in one link or in how links fit. Ikea's cost advantage lives in design and flat-pack logistics, not in cheap wood. A luxury brand's lives in marketing and service. The question is always: which activity do we do better, and can they copy it? A copyable advantage lasts about as long as it takes to copy.

  Economies of scale and the learning curve
  Bigness helps in two plain ways. Fixed costs spread out: a factory costs the same whether it makes a thousand units or a million, so the average cost per unit falls with volume. And experience teaches: the more a firm makes, the better it gets. T. P. Wright noticed in 1936 that the labour per aeroplane fell by a steady percentage each time cumulative output doubled; consultancies later called this the experience curve.
  average cost = fixed cost ÷ volume + unit cost × (learning factor)learning factor = r^d, where r is the cost kept after each doubling of cumulative output and d the number of doublings
  
    Interactive · Scale and learning
    A firm with $F million of fixed cost makes Q million units. Its rival makes half as many and has had one doubling fewer of experience. Invented numbers; the shape is what matters.
    Fixed cost ($ million)
    Volume (million units)
    Cost kept after each doubling
    Doublings of experience (you)
    
    
  
  The catch: scale only helps if demand is large enough to use it. A giant plant sitting half-empty has higher average costs than a small full one. And a learning curve is not automatic; it comes from deliberate work, and rivals can also learn.

  Network effects and platforms
  Some goods get more valuable as more people use them. A telephone is useless alone and valuable with millions. Economists call this a network effect (a demand-side economy of scale): it raises what buyers are willing to pay, rather than lowering the seller's cost. Direct effects come from other users of the same thing (messaging, a phone network). Indirect or two-sided effects link groups: more drivers make a ride app better for riders, and more riders make it better for drivers. A platform is a business built to host such a market, and the economics of pricing two sides at once was set out by Jean-Charles Rochet and Jean Tirole (2003).
  Robert Metcalfe's law says a network's value grows with the square of its users, since n users have about n(n−1)/2 possible connections. Andrew Odlyzko, Bob Briscoe and Benjamin Tilly argued in 2006 ("Metcalfe's Law is Wrong") that later users are worth less than early ones, so growth is nearer n × log n. Studies of Facebook and Tencent have fitted n², and a European internet-usage study found n² for small and n log n for large networks. The data haven't settled it; the main point survives either way.
  
    Interactive · Tipping point
    A toy model. Each person joins if their taste for the product, times its value (a base value plus the network's strength times the share who already use it), exceeds the price. Rounds repeat. Start below the critical mass and it dies; above it, it takes off.
    Network strength
    Price (as a share of top value)
    Starting share of users
    
    
  
  Three cautions. Tipping is not automatic: it needs network benefits to outweigh product differences and costs of switching or using several services at once. Some networks congest (more users can make it worse). And data advantages are often weaker and shorter-lived than classic network effects. The famous cases, such as VHS beating Betamax, are told with different details by different historians.

  Entry, price wars and the economist's view of competition
  Strategy books ask how to win. Industrial organization asks the opposite: how many rivals does it take for a market to behave competitively? The classic model is the French economist Antoine Cournot's (1838): firms choose how much to make, and the price falls as total output rises. With one firm you get the monopoly price. With two, a lower price. As the number of firms grows, price slides toward cost. Prices fall quickly at first and then slowly.
  
    Interactive · How many rivals?
    Demand is P = 100 − Q. Each of N identical firms chooses output; unit cost is c. HHI is the sum of squared market shares. The merge button removes one firm (two become one).
    Number of firms
    Unit cost
    
      Merge two firms
    
    
    
  
  Notice what the model leaves out, and what strategists add. Real firms also enter and exit, and an incumbent may threaten a price war to deter entry. Game theory (34/45) shows that a threat to fight is only believable if fighting is in the incumbent's own interest when the time comes; otherwise entrants call the bluff. Firms avoid price wars when they can: in the Cournot picture profit falls fast as rivals multiply, which is why firms prefer differentiating products to matching prices.

  Market power and antitrust
  Firms that keep profits by being better are doing what the system rewards. Firms that keep them by shutting rivals out are doing something else. Telling them apart is the job of competition law (called antitrust in the United States), and the regimes differ.
  
    RegimeMain rulesNotes
    United StatesSherman Act (1890), Clayton Act (1914), Hart-Scott-Rodino (1976)Sherman §2 bars monopolisation. Mergers are reviewed by the DOJ and FTC.
    European UnionArticles 101 (agreements) and 102 (abuse of dominance); Merger Regulation 139/2004Digital Markets Act applies since 2023 to designated "gatekeepers", fines up to 10% of worldwide turnover.
    ChinaAnti-Monopoly Law, in force since 2008Enforced since 2018 by the State Administration for Market Regulation (SAMR).
    IndiaCompetition Act, 2002Enforced by the Competition Commission of India.
  
  Recent examples, each dated. On 5 August 2024 a US federal judge, Amit Mehta, found that Google had illegally maintained a monopoly in general search; in a remedies decision of 2 September 2025 he declined to break the company up, barred exclusive contracts for search and ordered some data sharing, and both sides appealed. In April 2025 the European Commission imposed the first fines under the Digital Markets Act: €500 million on Apple and €200 million on Meta. In April 2021 China's regulator fined Alibaba about $2.8 billion, around 12% of its 2020 net profit, for anticompetitive practices.
  The tools differ in what they measure. A common first screen is the HHI (the widget above). The US 2010 guidelines treated a market above 2,500 as highly concentrated; the 2023 revision tightened the threshold, so check the current text. The deeper argument is about the goal. The consumer welfare standard, associated with Robert Bork, asks whether prices rise or output falls for buyers. Critics (sometimes called "neo-Brandeisian") say this is too narrow and misses power over workers, suppliers and politics. Defenders say it is the only standard that can be measured and tested. Most jurisdictions now treat consumer welfare as one factor among several.

  What a graduate just sees
  
    Industry structure explains a lot of profit before you look at the managers: score the five forces.
    An advantage is only as good as it is hard to copy.
    Your margin = rival's cost + your premium − your cost: an advantage is a gap.
    Average cost falls with volume, but only if demand fills the factory.
    Network effects create winners by tipping; critical mass is the whole game, and expectations decide it.
    In a market with few rivals, prices drift up; each extra rival helps, with diminishing force.
    A merger changes the concentration number, and regulators look at both the level and the change.
    "Dominant because better" and "dominant because blocked everyone else" can look the same from outside, and the law tries to tell them apart.
  

  Where it shows up next
  
    Pricing (34/61): once you know your cost and your rival's cost, pricing is choosing where in that gap to sit.
    Marketing Management (34/57): positioning is differentiation seen from the customer's side.
    Entrepreneurship (34/56): a new venture's first question is which force it can beat and for how long.
    In the world: why airlines and grocers earn thin margins, why a "winner takes most" app market forms, what a merger review asks, and every news story about a company being sued by a regulator.
  

  If you want to go deeper
  
    📘 Textbook. Robert Grant, Contemporary Strategy Analysis (Wiley). The standard strategy text for business students, with industry analysis, resources and cases.
    📘 Textbook. David Besanko, David Dranove, Mark Shanley and Scott Schaefer, Economics of Strategy (Wiley). The economist's version, with game theory throughout. For the pure theory, Jean Tirole's The Theory of Industrial Organization (MIT Press, 1988).
    📘 Gentler. Michael Porter, Competitive Strategy (1980), for the originals; Carl Shapiro and Hal Varian, Information Rules (1999), for networks and standards; Tim Wu, The Curse of Bigness (2018), for the antitrust story and the argument against consumer welfare.
    🌐 Free course. MIT OpenCourseWare 15.902 Strategic Management I (Fall 2006, Hax and Sahani), a graduate Sloan course on analytical methods and strategic practice.
  

  
    The takeaway
    Profit lasts where something stops rivals: high entry barriers, scale a rival cannot match, a network that tips, or a product buyers will not swap. Where nothing stops them, profit is competed down to cost. The economist adds a second question: when a firm's profit comes from blocking rivals rather than serving buyers, does society step in, and by what test?
    Try this: pick a business you use (a bank, a streaming service, a shop). Score its five forces from 1 to 5, then name the one thing that stops a rival from copying it. If you cannot name one, expect thin margins.
  

  
    Sources & Further Reading
    Sources. Margins: Aswath Damodaran, industry margins dataset, January 2026 (net margins: all firms 9.74%, software 25.49%, tobacco 26.65%, soft drinks 13.40%, air transport 2.51%, steel 1.93%, grocery and food retail 1.32%, autos and trucks 1.29%); the five-force ratings in the widget are the author's. Five forces: Porter, "How Competitive Forces Shape Strategy", Harvard Business Review (1979) and "The Five Competitive Forces That Shape Strategy" (2008); criticisms (Coyne and Subramaniam; Brandenburger and Nalebuff's complementors; the resource-based view) as summarised on Wikipedia. Generic strategies (1980) and the value chain (1985) are standard textbook attributions to Porter. Wright (1936) is the standard origin of the learning curve. Network effects: definitions and tipping caveats as summarised on Wikipedia; Metcalfe's law and its critics (Briscoe, Odlyzko and Tilly, IEEE Spectrum, 2006; Madureira et al. 2013; Zhang, Liu and Xu 2015) likewise; Rochet and Tirole (2003) on two-sided markets. Antitrust: statutes as listed on Wikipedia's competition-law article; United States v. Google (Mehta, 5 August 2024 liability ruling; 2 September 2025 remedies); Digital Markets Act (in force 1 November 2022, gatekeepers designated September 2023, first fines April 2025: Apple €500 million, Meta €200 million); Alibaba fine, SAMR, 9 April 2021 (as reported, $2.8 billion, about 12% of 2020 net profit); US HHI bands from the 2010 Horizontal Merger Guidelines (1,500 and 2,500). Every interactive number was computed in node: for example, with three firms and unit cost 20 the Cournot price is 40, each firm makes 20 and earns 400, HHI 3,333; merging two raises the price to 46.7; the tipping model with strength 1.5 and price 0.3 has a critical mass of 4.2% and a stable share of 79.1%.
    
      🌐 MIT OCW 15.902 Strategic Management I
      🌐 Damodaran: margins by industry
      📘 Besanko et al., Economics of Strategy
      📘 Tim Wu, The Curse of Bigness


Step 13. Product Innovation & Go-to-Market: Why Good Products Still Fail
------------------------------------------------------------------------
Why: Taking a new product from idea to market, and why most launches fail.
A new product has to be built, found by the right people, adopted in sequence, and sold at a cost that works. Most of what goes wrong happens after the engineering is done.

On 9 January 2007 a phone was announced; it went on sale on 29 June at $499 or $599 on a two-year contract. About 270,000 were sold in the first week, and in September the price of the larger model fell to $399. Six years earlier, on 3 December 2001, a two-wheeled personal transporter was unveiled after weeks of speculation that it could be as big as the personal computer. Its price was $5,000. Over its whole life about 140,000 were sold, and production of the original model ended in 2020.
  Both were clever machines. One changed an industry and the other found a niche. The subject of this piece is the thinking that tries to tell them apart before the money is spent: where products come from, how adoption spreads, how to know you have fit, how to launch, and why so many launches miss. The honest answer, as you will see, is that the field can describe the patterns better than it can predict the cases.

  
    The course in one box
    
      Who takes itMarketing majors in year 4, as an elective or capstone (marketing programmes such as Wharton's and Kellogg's offer new-product and innovation courses), and engineers and designers in product-development courses. Some business schools teach it only as an MBA elective, such as MIT Sloan's course 15.356, How to Develop Breakthrough Products and Services.
      What it is forTaking a new product from idea to the first paying customers, and deciding with evidence whether to go on.
      Comes afterMarketing Management (34/57) and Marketing Research & Analytics (34/59).
      Leads toIt closes the Marketing major and feeds Entrepreneurship (34/56). Close neighbours: Pricing (34/61), Brand Management (34/60), Digital Marketing & Advertising (34/62).
      What this piece gives youYou will not be able to run a launch. You will know the models people use (jobs, diffusion, fit, unit economics), which of their claims hold up, and why the famous "95% of new products fail" figure should be handled with care. Easier reads: The Smallest Possible Version (04/01), How to Talk to Customers (04/05).
    
  

  Where products come from, and the job they are hired for
  New products come from three directions. Technology push: a lab makes something possible and a company hunts for a use. Market pull: customers have a problem and someone builds the answer. And users themselves: MIT's Eric von Hippel showed that people who use a product in demanding ways often build the first version, and that the "lead users" (whose needs run ahead of the market) are a rich source of ideas. In one of his multi-industry studies, lead users were the pioneering developers of about half of the innovations judged most important.
  The best-known lens for the pull side is jobs to be done. The idea is that customers do not buy a product; they "hire" it to make progress in a situation. Clayton Christensen's milkshake story: a fast-food chain found many milkshakes sold early in the morning to solo commuters, hired for a job (make a boring drive less dull and keep hunger away until lunch), so the competitor was a banana or a bagel, not another milkshake. The method was set out in Competing Against Luck (2016, Christensen, Hall, Dillon and Duncan). The older line that people want "a quarter-inch hole, not a drill" is usually credited to Theodore Levitt, though its exact origin is murky.
  The critics have a point, and the field is split. There are rival schools: Christensen's, and Tony Ulwick's "outcome-driven innovation", which defines a job as a measurable set of desired outcomes. "Job" is defined differently across them, the milkshake case is one consulting story rather than a test, and the framework is hard to falsify. Christensen's earlier theory of disruption has had similar scrutiny: Jill Lepore's 2014 New Yorker essay questioned its evidence and its reliance on selected cases. Treat jobs to be done as a useful question ("what is this person trying to get done?") rather than a proven method.

  Diffusion and the Bass curve: how a good idea spreads
  Everett Rogers's Diffusion of Innovations (1962; 5th edition 2003) gathered hundreds of studies, beginning with Ryan and Gross's 1943 study of hybrid seed corn among Iowa farmers, which took over ten years to spread even though the corn paid for itself. Rogers sorted adopters by when they adopt: innovators (first 2.5%), early adopters (next 13.5%), early majority (34%), late majority (34%) and laggards (16%). The cut-offs come from slicing a bell curve; they are a convention, not a law of nature.
  Frank Bass turned the picture into one equation in 1969. At any moment, the chance a non-adopter buys has two parts: an outside push that does not depend on anyone else (p: advertising, curiosity) and an inside pull that grows with how many people already have it (q: word of mouth, imitation).
  dF/dt = (1 − F)(p + q F)share who adopt this instant = (share still to go) × (outside push + pull from those already in)
  Across many consumer products, studies report an average p of about 0.03 and q of about 0.38 per year. With those, sales peak at t* = ln(q/p) ÷ (p+q), about 6.2 years. Tune it yourself.
  
    Interactive · Tune an adoption curve
    
      Typical (p 0.03, q 0.38)
      Viral word of mouth
      Ad-driven, no buzz
    
    Outside push p (per year)
    Word-of-mouth pull q (per year)
    
    
    Bass model for a market of 1,000,000 potential buyers, 15 years. Bars: new adopters each year. Line: share adopted so far. Illustrative.
  
  Notice the lesson in the second preset: with weak outside push, a strongly social product starts slowly and then explodes. With only outside push, sales are highest on day one and fall steadily. The shape of the curve tells you what to spend on: reach in the first case, referral in the second.

  Crossing the chasm: when early adopters are not the market
  Geoffrey Moore's Crossing the Chasm (1991) says the Rogers curve has a crack in it. Early adopters (visionaries who will tolerate a rough product for a big payoff) buy for different reasons from the early majority (pragmatists who want a complete solution and references from other pragmatists in their own industry). Praise from enthusiasts does not move them. Moore's advice is to pick one narrow niche, win it completely with a full solution, and use its references to cross, a bit like securing a beachhead before the main landing.
  It is a famous idea with a contested evidence base. Rogers himself rejected it, writing that past research showed no support for a chasm between adopter categories. Goldenberg, Libai and Muller (2002) found a "saddle" (a dip in sales after the early peak) in a minority of consumer products, which they attribute to weak communication between segments, not to a universal gap. The toy model below has two groups, early adopters (16%) and a mainstream that mostly copies its own kind. It is mine, not Moore's.
  
    Interactive · Is there a chasm?
    
      Mainstream listens to early adopters
      Chasm
      Beachhead niche
      Stuck
    
    Mainstream copies early adopters
    Mainstream's own word of mouth
    
    
    Two-segment Bass-style model: early adopters (16%) p 0.25, q 0.5; mainstream (84%) p 0.004, own q as set, plus the chosen pull from early adopters. Illustrative, not data.
  

  Product–market fit: how you know you have something
  "Product–market fit" means a product that satisfies a strong demand in a market, so that customers pull it out of you. Andy Rachleff of Benchmark named the idea; Marc Andreessen popularised it in the mid-2000s with the claim that it was the only thing that mattered for a startup. How do you see it? Sean Ellis, who looked at many startups, proposed one question: how would you feel if you could no longer use this product? If at least 40% of users answer "very disappointed", you are probably there. The 40% is a rule of thumb from experience, not a law.
  The trouble is noise. A survey of 30 people cannot tell 40% from 25%. And the people who answer are not everyone: users who stuck around, who were willing to fill in a form. Both biases push the number up. Use the tool to see how wide the doubt is.
  
    Interactive · Does 40% really clear the bar?
    People surveyed
    Say "very disappointed"
    True share in the whole user base
    Run the survey again
    
    95% Wilson interval for a proportion. The re-run draws random samples from the "true share" you set: each press is a fresh simulated survey of the same people.
  

  Launch planning and channels: the economics of getting found
  A launch is a set of choices: who the first customers are, what promise you make them, where they will hear it, what you charge and when you go. A cheap way to learn is a staged launch: a small group first, then a wider one, so mistakes are small. For channels, Gabriel Weinberg and Justin Mares's Traction (2015) lists 19 (search ads, content, partnerships, sales, referrals and more) and suggests a "bullseye": brainstorm all, test three cheaply, then focus on the one that works.
  To compare channels you need two numbers. Customer acquisition cost (CAC) is what you spend to win one paying customer. Lifetime value (LTV) is the profit that customer brings before leaving. Investors in subscription businesses often quote rules of thumb: an LTV of at least three times CAC, and a payback of under a year. They are conventions, not laws, and they vary by industry.
  
    Interactive · A launch channel's unit economics
    
      Paid ads
      Content and search
      Outbound sales
      Referrals
    
    Spend on the channel
    Cost per visitor
    Visitors who sign up
    Sign-ups who pay
    Price per month
    Gross margin
    Months a customer stays
    
    Presets are invented, illustrative numbers, not benchmarks for any real channel.
  

  Why most launches fail: a number nobody can source
  You will often read that "95% of new products fail", usually credited to Christensen, or that 80% or 90% do. Be careful. In a 2013 review in the Journal of Product Innovation Management, George Castellion and Stephen Markham traced the best-known failure rates and reported that the very high figures lack solid support, and that careful surveys of companies' own launches imply a much lower rate. I could not retrieve the paper's exact numbers while writing, so no number is quoted here.
  The deeper problem is that "fail" means many things:
  
    Definition of failureGives a rate that is
    Missed the internal sales targetHigh: targets are often hopeful
    Pulled from the market within a few yearsModerate
    Never recovered its development costLower than the first, higher than you think
  
  A different measure with real data: the US Bureau of Labor Statistics follows every private-sector business establishment (new products are not tracked separately). About 47% of establishments that opened in the year to March 2021 were still open five years later, and about 34% of those opened in the year to March 2016 survived ten years. That is about business survival in general, not the same as product failure, but it fits the same pattern: failure is common, though not 95%.

  Launches and flops, told straight
  
    ProductWhat the record showsWhat the usual story gets wrong
    Edsel (Ford), 1957Launched 4 Sep 1957, cancelled 19 Nov 1959; about 116,000 sold, under half the break-even point."Research killed it" is too simple: a recession, prices overlapping Mercury, a divisive grille and quality complaints all contributed.
    New Coke, 1985Launched 23 Apr; "Classic" back 11 Jul (79 days). The new taste beat both Coke and Pepsi in blind tastings.Tastings missed attachment to the old brand. Executives denied a deliberate plot ("we're not that dumb, and we're not that smart").
    Segway, 2001$5,000 at launch; about 140,000 sold in its lifetime; the original model ended 2020.The hype, not the machine, was the failure; its descendants live on as scooters and in tourism and security.
    Google Glass, 2013$1,500 for "Explorers"; consumer sales ended Jan 2015; business editions ended 2023.It was a social acceptance problem as much as a technical one.
    Quibi, 2020Raised $1.75 billion; launched 6 Apr; about 500,000 subscribers; shut down by Dec 2020.Its founder blamed the pandemic; others point to the product and the content.
  

  What a graduate just sees
  
    People hire products for jobs. The real competitor may not be in your category.
    Adoption is social: after the first few, a product spreads through those who already have it.
    Early adopters can mislead you; the mainstream wants references from its own kind.
    A fit survey on 30 people is noise; ask who answered and how many.
    Every channel has a CAC; if LTV is not comfortably above it, growth makes the loss bigger.
    "Most products fail" is true and its famous percentage is not well supported.
    Every flop story has a tidier version and a messier documented one.
  

  Where it shows up next
  
    Entrepreneurship (34/56): a new venture is this subject with no budget or brand yet.
    Pricing (34/61) and Brand Management (34/60): the launch price and name are the first decisions of both.
    Marketing Research & Analytics (34/59): the survey and A/B tests that decide whether to go on.
    In the world: product managers, brand managers, startup founders, and the engineers who sit with them.
  

  If you want to go deeper
  
    📘 Textbook. Karl Ulrich and Steven Eppinger, Product Design and Development (McGraw-Hill). The standard engineering and design text.
    📘 Textbook. Philip Kotler and Kevin Lane Keller, Marketing Management (Pearson). The chapters on developing and launching new offerings.
    📘 The classics. Everett Rogers, Diffusion of Innovations (5th ed., 2003); Geoffrey Moore, Crossing the Chasm (3rd ed., 2014).
    📘 Gentler. Gabriel Weinberg and Justin Mares, Traction (2015); Christensen et al., Competing Against Luck (2016).
    🌐 Free course. MIT OpenCourseWare 15.356 How to Develop Breakthrough Products and Services (Spring 2012), and 15.810 Marketing Management.
  

  
    The takeaway
    A product succeeds when it does a job someone cares about, reaches them through the right people in the right order, and costs less to win than it earns. The models (jobs, Bass, the chasm, fit, unit economics) are good questions rather than forecasts, and the famous failure rates and flop stories deserve a second look.
    Try this: pick a product you started using in the last year. Write down the job you hired it for, who told you about it, and what you stopped using.
  

  
    Sources & Further Reading
    Sources. Bass: Frank Bass, "A New Product Growth Model for Consumer Durables", Management Science 15(5), 1969; closed form F(t) = (1 − e−(p+q)t) ÷ (1 + (q/p) e−(p+q)t), peak t* = ln(q/p) ÷ (p+q); averages p 0.03 and q 0.38 from Mahajan, Muller and Bass (1995) as summarised in standard references. Rogers, Diffusion of Innovations (1962; 5th ed. 2003); Ryan and Gross (1943) hybrid corn in Iowa; adopter shares 2.5/13.5/34/34/16 are Rogers's standard categories. Moore, Crossing the Chasm (1991, revised 1999 and 2014); Rogers's rejection of a chasm; Goldenberg, Libai and Muller, "Riding the saddle", Journal of Marketing (2002), summarised from memory, share of products with a saddle not stated. Von Hippel lead-user work (1986; the nine-industry 54.4% figure is as reported in secondary sources). Rachleff and Andreessen on product–market fit; Sean Ellis 40% test (rule of thumb). Lepore, The New Yorker (2014). Castellion and Markham (2013), Journal of Product Innovation Management: cited from memory, not re-read. BLS Business Employment Dynamics, establishment survival (accessed 2026): 47.2% at five years for the year to March 2021, 34.2% at ten years for the year to March 2016. iPhone: announced 9 January 2007, released 29 June 2007, $499/$599, price cut 5 September 2007, about 270,000 sold in the first week. Edsel, New Coke, Segway, Google Glass and Quibi dates and figures from standard references (Wikipedia summaries of contemporary press and company statements). Every number in the interactives was computed in node; the toy two-segment model is the author's own.
    
      🌐 MIT OCW 15.356
      📘 Rogers, Diffusion of Innovations
      📘 Moore, Crossing the Chasm
      📘 Ulrich & Eppinger, Product Design and Development


Step 14. Econometrics: How to Find What Caused What
---------------------------------------------------
Why: Proving what a campaign actually caused.
Correlation is cheap. Econometrics is the craft of getting from "these move together" to "this made that happen", and of knowing how badly you could be fooled on the way.

A city opens a new tram line, and a year later house prices along it have risen. Did the tram do it? Maybe. Or maybe the city built the line where it already expected the area to boom. Or a new university opened nearby the same year. Every headline that says "X is linked to Y" hides this question, and most of the time the headline cannot answer it.
  An economist sees the same pattern in a minimum-wage law, a school-class size, a tax change, a drug trial. The data show a link. The job is to say what the link would have been if nothing else had been going on. That is econometrics: statistics built for questions where you usually cannot run the experiment, and where the people in your data chose their own circumstances.

  
    The course in one box
    
      Who takes itEconomics majors, usually in year 2 or 3, straight after a first statistics course. Public policy, political science, finance and business-analytics programmes teach close cousins. Programmes such as MIT's Course 14-1, the LSE and Chicago all require it.
      What it is forEstimating the effect of one thing on another from messy real-world data, with an honest statement of how wrong the estimate might be.
      Comes afterProbability & Statistics (34/05).
      Leads toDevelopment Economics (34/48), and it is the working tool in Behavioral Economics (34/46), Marketing Research & Analytics (34/59) and Machine Learning (34/32).
      What this piece gives youYou will not be able to run a regression package after this. You will see why a coefficient can lie, what a clever study does to stop it lying, and why a "significant" result is weaker evidence than it sounds. It goes further than Probability & Statistics (34/05), whose last section stops at "correlation is not causation"; here is what to do about it. The library's How to Spot Bullshit (17/03) is a gentler start. All datasets here are invented and labelled illustrative.
    
  

  Regression is the best-fitting line
  Take twelve students: hours studied, and the score each got. Plot them and the dots drift upward. A regression draws the one straight line that sits closest to all the dots. "Closest" has a precise meaning: for each dot, measure the vertical gap to the line (the residual), square it, and add up the squares. The line with the smallest total is the ordinary least squares (OLS) line. Squaring means one big miss counts far more than several small ones; the picture shows each squared miss as a literal square.
  score = a + b × hours + errorb is the slope: the score gap between two students who differ by one hour
  
    Interactive · Fit the line yourself (illustrative data)
    Slope b (points per hour)
      
    Intercept a (score at 0 hours)
      
    Snap to the best line
    
    
    Twelve invented students. The best line is found by solving the least-squares equations; no other line has a smaller sum of squared misses.
  
  The best line here is about 47 + 3.5 × hours (the readout gives it exactly). The coefficient b says: students who studied an hour more scored about 3.5 points higher. Notice what that sentence does not say. It compares different students. It does not say that making one student study an extra hour would raise her score by 3.5. That gap, between a comparison and a cause, is the rest of the course.

  Omitted variables: add the confounder, watch the effect change
  Suppose you regress wages on years of schooling in a big survey and find that each extra year goes with higher pay. Is that what schooling does? People with more drive or talent may both stay in school longer and earn more anyway. If so, the schooling coefficient is also soaking up the talent effect. This is omitted-variable bias, and it has a tidy rule:
  bias ≈ (effect of the omitted factor on the outcome) × (how strongly it moves with your variable)if either is zero, there is no bias
  In the lab below, the true effect of one more year of schooling is exactly 3 (thousand dollars). Talent also raises pay, and a dial sets how strongly talent pushes people toward more school. Switch the model between ignoring talent and including it.
  
    Interactive · The hidden confounder (illustrative simulation, 1,500 people)
    How strongly talent pushes people to stay in school
      
    
      Ignore talent
      Control for talent
      Randomise schooling
    
    
    above-average talent   below-average talent
    
    True schooling effect = 3.0; talent adds 8 per standard deviation. Wage = 10 + 3·school + 8·talent + noise. "Randomise" means a coin-flip-like process assigns schooling regardless of talent.
  
  Without JavaScript: with talent pushing schooling at strength 1.2, ignoring talent gives a schooling coefficient near 4.7 instead of the true 3.0; controlling for talent recovers about 3.0; randomising schooling also recovers about 3.0.
  Ignoring talent at strength 1.2 gives about 4.7, over half again too large. Control for talent and it falls to about 3. The cost of this fix: you must measure the confounder. Most real confounders (motivation, family connections, how well-run the firm is) sit in no dataset. That is why economists spent decades arguing about what regressions could prove.

  Randomise: the coin that cuts every hidden link
  Press the third button. When schooling is assigned by a process unrelated to talent, the link between them is broken, and even the naive regression that ignores talent lands on 3. A coin flip is independent of everything: motivation, family, luck, things nobody thought to measure. That is why a randomised experiment is the gold standard, and why 2019's Nobel prize in economics went to Abhijit Banerjee, Esther Duflo and Michael Kremer for experimental approaches to reducing poverty.
  But economists rarely get to flip the coin: you cannot randomly assign years of schooling, a recession or a minimum wage. The next three ideas are ways to find a coin that nature or a rule has already flipped. Together they are called the credibility revolution, a phrase from Joshua Angrist and Jörn-Steffen Pischke's 2010 essay, answering Edward Leamer's 1983 complaint that econometrics needed to "take the con out" of itself.

  Difference-in-differences: borrow the other group's trend
  In 1992 New Jersey raised its minimum wage and neighbouring Pennsylvania did not. David Card and Alan Krueger surveyed fast-food restaurants in both states before and after (American Economic Review, 1994). Looking only at New Jersey before and after would confuse the law with anything else changing that year. Instead they used Pennsylvania as the stand-in for "what would have happened anyway". Their finding, that employment did not fall in New Jersey, contradicted the textbook prediction at the time. It has been disputed ever since: Neumark and Wascher re-analysed it with payroll records and found job losses, and the field's overall reading is still argued over. The design is not in dispute. Card shared the 2021 Nobel prize in part for it.
  effect = (treated after − treated before) − (comparison after − comparison before)subtract the comparison group's change to remove the common trend
  
    Interactive · Difference-in-differences (illustrative numbers)
    Common trend (hits both groups)
      
    True effect of the policy
      
    Treated group was already drifting differently
      
    
    
    Treated group starts at 20, comparison at 24 (invented units). Dashed line: where the treated group would have gone without the policy, if it followed the comparison's trend.
  
  With a common trend of +1.5 and no true effect, the treated group still rises by 1.5, so a naive "before and after" would claim a policy success. The difference-in-differences correctly says zero. The design's weak point is the third slider: it assumes the two groups would have moved in parallel without the policy. If the treated group was already on a different path, the estimate is off by exactly that gap. Good papers show the two lines tracking each other for years before the policy.

  Instruments and cutoffs: two more ways to find a coin
  An instrumental variable is something that nudges the thing you care about but has no other route to the outcome. Angrist (1990) studied whether military service in the Vietnam era changed later earnings. Men who chose to serve differ from those who did not, so a plain comparison is confounded. But the draft lottery assigned numbers by birthday at random, and a low number made service much more likely. The lottery number is a coin that pushes people into uniform. Its effect on earnings, divided by its effect on service, isolates the effect of service for the people who served because of the lottery. His study found that service lowered earnings for the veterans concerned.
  effect of service = (earnings gap by lottery group) ÷ (service gap by lottery group)the "Wald" ratio: reduced form over first stage
  A worked number, invented to show the arithmetic: suppose 35% of low-number men served and 20% of high-number men, a first stage of 0.15; and low-number men earned $300 less a year on average. Then the estimated effect of serving is −300 ÷ 0.15 = −$2,000 a year. Dividing by a small first stage blows up noise, which is why weak instruments are dangerous.
  A regression discontinuity uses a cut-off. Israeli school classes are capped at 40 pupils under a rule attributed to the twelfth-century scholar Maimonides, so a school with 40 pupils has one class and a school with 41 splits into two small classes of about 20. Angrist and Lavy (1999) compared schools just either side of those thresholds, where enrolment is close to a coin flip, to estimate what class size does to test scores. Slide the window below: narrow windows are fairer but noisier.
  
    Interactive · Regression discontinuity (illustrative data)
    Window around the cut-off (± points)
      
    True jump at the cut-off
      
    
    
    400 invented students; those scoring 60 or more get a support programme. Outcome rises 0.3 per score point anyway, so just comparing everyone above with everyone below overstates the jump.
  

  Standard errors: how much would the answer wobble?
  Every estimate comes from one sample, and a different sample would have given a different number. The standard error is the typical size of that wobble. For a regression slope it shrinks with the square root of the sample size, grows with the noise around the line, and shrinks when x varies a lot:
  SE(b) ≈ (noise around the line) ÷ ( spread of x × √n )four times the data halves the wobble
  The rule of thumb: an estimate more than about two standard errors from zero is called statistically significant at the 5% level. Significant means "hard to produce by sampling wobble alone if the true effect were zero". It does not mean big or important, and it does not mean true. Run 300 studies of a real but modest effect below and see who gets stars.
  
    Interactive · Run 300 studies (illustrative simulation)
    Sample size per study
      
    True slope
      
    Run 300 new studies
    
    significant at 5%   not significant
    
  
  At n = 50 and a true slope of 0.3, a little over half of the studies reach significance, and those average about 0.40, a third above the truth, because only the lucky overestimates clear the bar. Drop to n = 10 and a true slope of 0.05 and the few stars (about 8%) average roughly 0.7, fourteen times too big. Published results are mostly the significant ones, so they exaggerate. That is one reason newer replications so often find smaller effects.

  P-hacking: test enough things and something will glow
  A test at the 5% level is wrong by chance 1 time in 20 when nothing is going on. Test 20 unrelated outcomes and the chance that at least one glows is 1 − 0.9520 = 64%. A researcher who tries several outcomes, subgroups or control sets and reports the one that "worked" is p-hacking, whether on purpose or just by wandering a garden of forking paths. Simmons, Nelson and Simonsohn (2011) showed how easy it is: using flexible, ordinary-looking choices they "found" that listening to a song about being older made students younger by about a year and a half. Economists are not exempt: Brodeur and co-authors (2016) found an unnatural pile-up of results just below the 0.05 line in top economics journals.
  
    Interactive · A study of nothing (illustrative simulation)
    Outcomes you test (all truly unaffected)
      
    Run one studyRun 500 studies
    
    
    Each outcome compares two groups of 50 drawn from the same distribution, so every "effect" is noise. A cell lights up when |t| passes the 5% cut-off (1.98).
  
  Defences: decide the analysis before seeing the data (pre-registration), report every test you ran, adjust the threshold when you run many, and above all prefer designs, like the ones above, where the answer does not depend on which controls you picked.

  What a graduate just sees
  
    A regression coefficient is a comparison between different people, not a lever you can pull.
    Before believing "X raises Y", ask what else moves with X and also moves Y. The bias is its effect times its link to X.
    Adding controls can only fix confounders you measured. A coin flip fixes the ones you did not.
    Every natural-experiment study is a hunt for a coin: a lottery, a border, a cut-off, a law that changed for some and not others.
    Difference-in-differences needs parallel trends; ask to see the years before the policy.
    "Significant" means hard to get from wobble alone. It says nothing about size, and a small-sample star probably overstates.
    If you can run twenty tests, one will light up. Ask how many were run.
  

  Where it shows up next
  
    Development Economics (34/48): randomised trials and natural experiments are how economists test what actually helps poor countries.
    Behavioral Economics (34/46): the replication record in that piece depends on this one, from effect sizes to p-hacking.
    Marketing Research & Analytics (34/59): "did the campaign cause the sales, or did summer?" is difference-in-differences.
    Machine Learning (34/32): prediction asks "what will happen"; econometrics asks "what would happen if I changed it". They need each other.
    In the world: central banks and tax agencies, tech companies' A/B test teams, public-health trials, and every think-tank report that says "causes".
  

  If you want to go deeper
  
    📘 Textbook. James Stock and Mark Watson, Introduction to Econometrics (Pearson). A standard undergraduate text, strong on causal thinking and applications.
    📘 Textbook. Jeffrey Wooldridge, Introductory Econometrics: A Modern Approach (Cengage). The other standard choice, more algebra and more worked data.
    📘 Gentler. Joshua Angrist and Jörn-Steffen Pischke, Mastering 'Metrics: The Path from Cause to Effect (Princeton, 2015). Written for people who are not specialists, built around exactly the designs above.
    🌐 Free course. MIT OpenCourseWare 14.32 Econometrics (Spring 2007, Joshua Angrist): undergraduate course with lecture notes and problem sets.
  

  
    The takeaway
    A regression gives you a comparison, and a comparison is only a cause if nothing else differs between the groups. Econometrics is a toolkit for making that true or nearly true (randomise, difference out a trend, use a lottery, zoom in on a cut-off), plus the humility to report how much your number would wobble and how many ways you tried.
    Try this: the next time a headline says "X linked to Y", name one confounder that could drive both, and ask what coin flip, border or cut-off could separate them.
  

  
    Sources & Further Reading
    Sources. Least squares, omitted-variable bias, standard errors and the Wald estimator are standard textbook material (Stock & Watson; Wooldridge; Angrist & Pischke). Card & Krueger, "Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania," American Economic Review 84(4), 1994; the Nobel Prize in Economic Sciences 2021 went to David Card (half) and Joshua Angrist and Guido Imbens (half, "for their methodological contributions to the analysis of causal relationships"). Angrist, "Lifetime Earnings and the Vietnam Era Draft Lottery," 1990; Angrist & Lavy, "Using Maimonides' Rule to Estimate the Effect of Class Size on Scholastic Achievement," 1999; regression discontinuity was introduced by Thistlethwaite & Campbell in 1960. Angrist & Pischke, "The Credibility Revolution in Empirical Economics," Journal of Economic Perspectives 2010; Leamer, "Let's Take the Con out of Econometrics," 1983. Simmons, Nelson & Simonsohn, "False-Positive Psychology," 2011; Brodeur, Lé, Sangnier & Zylberberg, "Star Wars: The Empirics Strike Back," 2016. The 2019 economics Nobel went to Banerjee, Duflo and Kremer. Every number an interactive shows was computed in node from the code on this page: the best line is 47.0 + 3.53 × hours; omitted-variable bias at strength 1.2 is 8 × 1.2 ÷ (1.2² + 4) = 1.76 in theory (1.67 in this sample); 1 − 0.9520 = 64.2%. All datasets are invented.
    
      🌐 MIT OCW 14.32 Econometrics
      📘 Angrist & Pischke, Mastering 'Metrics
      📘 Stock & Watson, Introduction to Econometrics


Step 15. Machine Learning: Programs That Write Their Own Rules
--------------------------------------------------------------
Why: Recommendations, targeting and forecasting run on machine learning.
Nobody wrote the rule that tells a spam message from your mother's email. A program found it by looking at examples, and everything in the field is about whether you can trust what it found.

Your phone groups your photos by person. Your inbox hides the lottery scams. A bank flags a card payment at 3 a.m. from another continent. Try to write the rules for any of these by hand and you hit a wall within an hour: what, exactly, makes a face a face, or a message spam?
        Machine learning gives up on writing the rules. Instead you show a program thousands of labelled examples and let it adjust itself until it gets them right. This is powerful, and it quietly changes the questions you have to ask. The hard part is no longer "what is the rule?" but "how do I know the rule it found will work tomorrow, on people it has never seen?"
        That question, and a handful of simple tools for answering it, are what a graduate of a machine-learning course keeps for life.

        
            The course in one box
            Who takes it: Computer Science majors, usually in year 4; also statistics, data science and engineering students.
            What it's for: building programs that learn their behaviour from data, and judging honestly whether they work.
            Comes after: Linear Algebra (34/04), Probability & Statistics (34/05), Multivariable Calculus (34/02) and Algorithms (34/24).
            Leads to: no later subject in this series lists it as a prerequisite; it feeds research and careers in AI, and touches Computer Security (34/33), where learning systems are both tools and targets, and Databases (34/28), where the training data lives.
            Honest line: you won't train a modern language model after this; you will understand the training loop all of them share. For the AI story and a tour of language models, read Artificial Intelligence (29/01) and How Large Language Models Actually Work (29/02) first; this piece goes underneath them.
        

        Learning from data, and the data you hold back
        Take a toy problem. Each email gets a "suspicious words" score, and you must decide spam or not. You could write a rule by hand: spam if the score is above 3. But why 3? Better: try every threshold on a set of labelled examples and keep the one that gets the most right. That, in miniature, is training: a rule with an adjustable knob, a way to score the rule on data, and a search for the best setting of the knob.
        Here is the trap. The setting that fits the examples you used will almost always look better on those examples than on fresh ones, because you tuned it to them. So practitioners split their data three ways:
        
            Training set: the examples the model adjusts itself on.
            Validation set: held back to compare choices (which model, which settings).
            Test set: touched once, at the very end, to estimate how it will do in the real world. If you keep tuning against it, it stops being a test.
        

        
            Try it: pick the threshold, then meet fresh data
            Filled dots are 30 training emails (orange = spam, blue = not spam); hollow squares are 20 held-back ones the rule never saw. All the data are made up for this demo. The best threshold on the training emails is 3.3, which gets 93% of them right, but only 85% of the 20 held-back emails.
            
                
                Threshold: spam if score > 3.0
                    
                Let the data choose
                
            
        

        Linear regression and gradient descent
        Many problems are about predicting a number. The simplest model is a straight line: prediction = w × input + b. The two knobs are the slope w and the intercept b. To score a line, measure how far it misses each point, square the misses so that over and under both count, and average them: the mean squared error (MSE). The method of least squares was first published by Adrien-Marie Legendre in 1805; Carl Friedrich Gauss claimed to have used it earlier, which started a priority dispute.
        For a line there is a formula for the best w and b. For huge models there is not, and we use the idea that carries the whole field: gradient descent. Stand on the error surface, feel which way is steepest downhill (the gradient, from Multivariable Calculus 34/02), and take a small step that way. Repeat. Augustin-Louis Cauchy suggested the method in 1847.
        new w = w − η × (slope of the error with respect to w)    same for bη ("eta") is the learning rate, the size of each step
        The learning rate is a real trade-off: too small and you crawl; too large and you overshoot the valley and bounce out of it. Try it. The 12 points below are made up around a line; the least-squares answer is w = 2.465, b = 8.289, with MSE 2.58.

        
            Try it: fit a line by rolling downhill
            Start from w = 0, b = 0 and press Step. Choose a learning rate. With η = 0.01, 200 steps reach w = 2.89, b = 5.79 and MSE 4.37, still on the way down; with η = 0.04 the steps are too big and the error explodes instead of shrinking.
            
                
                
                    η = 0.001
                    η = 0.01
                    η = 0.03
                    η = 0.04
                
                StepRun 20 stepsReset
                
            
        

        Overfitting and underfitting
        A line is a rigid model. A more flexible one, say a polynomial with more bends, can follow the data more closely. Flexibility is a dial, and both ends of it fail:
        
            Underfitting: the model is too simple to capture the pattern. It is wrong on the training data and wrong on new data.
            Overfitting: the model is so flexible that it memorises the noise in the training data. It is nearly perfect there and bad on new data.
        
        This is why the validation set exists. Training error only ever goes down as you add flexibility. Validation error goes down, then comes back up, and the bottom of that U is where you stop.
        Training error tells you how well the model remembers. Validation error tells you how well it learned.

        
            Try it: turn the complexity dial
            The true pattern is a smooth wave; the 12 filled dots are noisy samples of it, and the 40 hollow dots are fresh noisy samples used only to judge. The curve is a polynomial fitted to the filled dots. A degree-1 line misses the wave (errors 0.54 and 0.43), degree 3 does best on fresh data (0.28 train, 0.35 validation), degree 9 hugs the 12 dots (0.14) but is far worse on fresh ones (1.02).
            
                
                Polynomial degree: 3
                    
                
            
        

        Classification and neural networks as layered functions
        Classification predicts a label rather than a number: spam or not, which digit, which animal. A simple classifier computes a weighted sum of the inputs and squashes it into a probability between 0 and 1, and its decision boundary is a straight line (or plane). That is enough when the classes can be separated by a line.
        Some problems cannot. The classic example is XOR: the answer is 1 when exactly one of the two inputs is 1. No single straight line separates the four cases correctly. In the late 1960s Marvin Minsky and Seymour Papert's book Perceptrons analysed exactly this kind of limit for single-layer networks, and is often credited with helping to cool interest in neural networks for years (historians debate how much).
        The escape is to stack models. A neural network is a chain of simple functions: each layer takes a weighted sum of its inputs, applies a bend (a non-linear function such as tanh), and hands the result to the next layer. One layer draws a line; a layer built on top of lines can carve out regions that no single line can. Training a network means finding good weights, and the tool is backpropagation, which in words goes:
        
            Run an example forward through the layers and get a prediction.
            Measure how wrong it was.
            Walk backwards, layer by layer, asking each weight "how much of this error was your fault?", using the chain rule from calculus.
            Nudge every weight a little in the direction that reduces the error. Repeat on the next example.
        
        The idea was derived several times; the 1986 Nature paper by David Rumelhart, Geoffrey Hinton and Ronald Williams is the one that made it popular. Gradient descent plus backpropagation, run at scale, is still how neural networks are trained, including the large language models in 29/02.

        
            Try it: teach a network XOR
            Four training cases. Choose how many hidden units the network has, then slide through training steps (always the same starting weights). The picture shows the network's output over the plane. With no hidden layer, the best the model can do is output about 0.5 for every case (loss 0.693) and it stays stuck; with 4 hidden units, 500 steps already get all four right (loss 0.010).
            
                
                
                    No hidden layer
                    2 hidden units
                    4 hidden units
                
                Training steps: 500
                    
                
            
        

        Evaluation: accuracy is not enough
        "The model is 99% accurate" can be meaningless. Imagine a rare disease affecting 1% of 10,000 people. A model that always says "healthy" is 99% accurate and finds nobody. Count outcomes instead: true positives (TP), false positives (FP), true negatives (TN), false negatives (FN).
        precision = TP / (TP + FP): of those flagged, how many were rightrecall = TP / (TP + FN): of those who were sick, how many were found
        The two pull against each other. A model that flags only when it is certain has high precision and low recall; one that flags at the slightest hint has high recall and swamps you with false alarms. Where to set the threshold depends on what each error costs: missing a cancer is not like sending a good email to the spam folder. And when the real condition is rare, even a decent test produces mostly false alarms.

        
            Try it: a test with a threshold
            10,000 people. The test gives a score; sick people score on average 2 standard deviations higher than healthy ones (a decent test, invented for this demo). Move the threshold and the prevalence. At 1% prevalence and a threshold of 1.0, the test finds 84 of 100 sick people (recall 84%) but raises 1,571 false alarms, so only 5% of those flagged are actually sick.
            
                
                    1% are sick
                    10%
                    50%
                
                Flag when score > 1.0
                    
                
                
            
        

        Bias: the model learns the data you gave it
        A model has no opinion about fairness. It finds whatever pattern best fits the examples it was shown, including patterns in who is in the examples. If one group is a small part of the training data, the model's single rule is tuned to the majority and does worse on the rest.
        This is not hypothetical. In the 2018 Gender Shades study, Joy Buolamwini and Timnit Gebru found that the two benchmark face datasets they examined were 79.6% (IJB-A) and 86.2% (Adience) lighter-skinned subjects, and that three commercial gender-classification systems misclassified darker-skinned women at error rates of up to 34.7%, against at most 0.8% for lighter-skinned men.

        
            Try it: who is in the training data?
            Two groups, A and B, whose scores sit on different scales (an invented example). The model picks the one threshold that gets the most training cases right. Change how much of the training data is group B. With 2% of the data from group B the model is 84% accurate for A but 74% for B; at 50% both groups get 81%.
            
                Group B share of training data: 2%
                    
                
                
            
        
        More data from the same source does not fix this; different data, or a model checked group by group, can. Bias can also come from the labels (past human decisions the model copies) and from how a system is used. Checking performance for each group, not just on average, is the habit to keep.
        The same discipline applies to the newest systems. The Stanford HAI AI Index 2026 report describes a "jagged frontier": one model earned a gold medal at the International Mathematical Olympiad, yet the top model it tested read analog clocks correctly only 50.1% of the time. An average score, or a headline result, says little about how a model will do on the case in front of you; that has to be measured.

        What a graduate just sees
        
            Learning means tuning knobs to fit examples; the rule is found, not written.
            Performance on the data you trained on is not evidence. Only data held back is.
            Gradient descent is "step downhill"; the learning rate is the step size and it matters.
            Too simple and too flexible both fail; validation error has a U shape.
            A network is stacked simple functions, and backpropagation is the chain rule applied layer by layer.
            Accuracy hides rare cases; ask for precision, recall and the cost of each mistake.
            When a thing is rare, most alarms are false alarms.
            A model is only as fair as its data; check each group, not the average.
        

        Where it shows up next
        
            Linear Algebra (34/04) and Multivariable Calculus (34/02): weights are vectors and matrices, and the gradients are partial derivatives.
            Probability & Statistics (34/05): the train/test idea is the statistician's out-of-sample check, and precision is Bayes' theorem in disguise.
            Algorithms (34/24): training is optimisation, and its cost is why hardware like GPUs matters (see 29/12).
            Computer Security (34/33): learning systems can be fooled by crafted inputs and poisoned data.
            In the world: fraud screening, search ranking, recommendations, translation, medical imaging, and the chatbots in 29/02. As of the 2025 Stanford AI Index, the compute used to train notable models was reported to be doubling roughly every five months, which is why training is now an industrial activity.
        

        If you want to go deeper
        
            📘 An Introduction to Statistical Learning (James, Witten, Hastie, Tibshirani; 2nd edition 2021, with a Python edition in 2023): free from the authors, the standard first university text; clear and light on maths.
            📘 Deep Learning (Goodfellow, Bengio, Courville, MIT Press 2016): the standard graduate text on neural networks, free to read online.
            ▶️ Gentler: 3Blue1Brown's neural-networks video series builds the ideas above with pictures.
            🌐 Free courses: MIT OpenCourseWare 6.036 Introduction to Machine Learning and Stanford CS229.
        

        
            The takeaway
            Machine learning replaces hand-written rules with knobs tuned to examples, using gradient descent to find the settings. Everything else is about trust: hold data back to measure honestly, find the middle between too simple and too flexible, stack functions when a line is not enough, ask which errors matter, and check who is in the data.
            Try this: the next time you read "an AI that is 95% accurate", ask four questions. Accurate on what data? How rare is the thing it looks for? What does each kind of mistake cost? And has anyone checked it for the group you belong to?
        

        
            Sources: A.-M. Legendre (1805) and C. F. Gauss on least squares (priority dispute, as summarised in standard histories of statistics); A.-L. Cauchy (1847) on gradient descent; M. Minsky and S. Papert, Perceptrons, MIT Press, 1969; D. Rumelhart, G. Hinton and R. Williams, "Learning representations by back-propagating errors", Nature 323, 1986; J. Buolamwini and T. Gebru, "Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification", Proceedings of Machine Learning Research 81, 2018; Stanford HAI, AI Index Report 2025 (training compute doubling about every five months) and AI Index Report 2026 (the "jagged frontier": IMO gold, 50.1% on analog clocks); G. James, D. Witten, T. Hastie, R. Tibshirani, An Introduction to Statistical Learning. All datasets in the demos are invented, and every number they show was computed and cross-checked for this piece.
            Go deeper:
            📘 An Introduction to Statistical Learning (statlearning.com, free)
            📘 Deep Learning — Goodfellow, Bengio, Courville (deeplearningbook.org, free online)
            ▶️ 3Blue1Brown: Neural Networks
            🌐 MIT OCW 6.036 Introduction to Machine Learning


Step 16. Entrepreneurship: How to Be Wrong Cheaply
--------------------------------------------------
Why: A new venture is marketing with no budget and no brand yet.
Most new ventures fail, and the survivors mostly found out what customers wanted before they ran out of money. Entrepreneurship is the craft of finding out early, and of doing the arithmetic that says how long you have to learn.

Two people each have an idea for an app that helps restaurants cut food waste. One spends eight months and the savings from a previous job building it, and launches to silence. The other makes a one-page description, phones forty restaurant managers, and learns by Thursday that most of them already track waste on a whiteboard and would pay for something quite different: a tool that predicts tomorrow's orders. The second founder is not smarter. She has just found the cheapest place to be wrong.
  This is the capstone of the business degree because it uses everything else: accounting to know whether a customer pays back, finance to value the company and price its shares, marketing to find the customer, operations to deliver. It adds one thing that none of them has: working out what to build when nobody can tell you.

  
    The course in one box
    
      Who takes itBusiness students in the last year, MBAs (it is part of the first-year Required Curriculum at Harvard Business School), and engineers and scientists in entrepreneurship minors and courses such as MIT's 15.390 New Enterprises, taught by Bill Aulet and colleagues at the Sloan School.
      What it is forFinding a problem worth solving, testing the solution cheaply, building a business that pays back each customer, and funding it without losing control by accident.
      Comes afterAccounting (34/50), Corporate Finance (34/51) and Marketing Management (34/57).
      Leads toIt is a capstone. It feeds Product Innovation & Go-to-Market (34/63) and Pricing (34/61), and it is where the whole business major gets used at once.
      What this piece gives youYou will not be able to raise a funding round or run a company after this. You will see why most start-ups die of the same few causes, and you will be able to compute a customer's payback, a runway and a founder's slice after dilution. Easier first reads in the library: The Smallest Possible Version (04/01), Where Money Actually Comes From (04/02), How to Talk to Customers (04/05), and the Starting a Business module of The Default Course. This piece goes further. It describes how things work and is not legal, tax or investment advice.
    
  

  Start with a problem someone already pays to solve
  The usual mistake is to fall in love with a solution. The disciplined habit, taught almost everywhere that entrepreneurship is taught, is to start with a problem: who has it, how often, how badly, and what they do about it now. The last question matters most. A problem people already spend money, time or hacks on is real. A problem they only say is annoying may not be.
  Steve Blank, whose customer development method sits under much of modern start-up teaching, put the central point as "get out of the building": facts about customers are outside, not in your head. The practical craft is asking in a way that produces evidence rather than politeness. Rob Fitzpatrick's The Mom Test (2013) gives a good rule: ask about their life and past behaviour, not about your idea. A compliment is the cheapest thing a person can give you.
  
    What you hearHow much it tells you
    "Great idea, I'd definitely use that."Almost nothing. Costs them nothing to say.
    "You should add feature X."An opinion. Useful for ideas, weak as proof.
    "Last month I spent six hours doing this by hand."A real behaviour. Worth following up.
    A deposit, a pre-order, a signed letter of intent.The strongest signal: something was given up.
  

  Experiments are cheap; being wrong is expensive
  Eric Ries, who coined "lean start-up" in 2008 and published the book in 2011, drew on lean manufacturing and Blank's work to describe the loop: build the smallest thing that tests a belief, measure what customers do, learn, and decide whether to carry on or change direction. That smallest thing is the minimum viable product (MVP): not a bad product, but the cheapest experiment that teaches you the most. It can be a landing page, a hand-made service you pretend is software, or a video.
  The method has critics. Ben Horowitz wrote in 2010 that "running lean" can become an end in itself instead of a means to winning a market. Teppo Felin, Alfonso Gambardella, Scott Stern and Todd Zenger argued that copying factory-style iteration yields only incremental ideas, and that bold start-ups need a theory, not just A/B tests. Others note that some products, such as medical devices or hardware, cannot be cheaply tested. In its favour: a randomised trial of Italian early-stage ventures (Camuffo, Cordova, Gambardella and Spina, Management Science, 2020), where entrepreneurs were taught to treat their idea as a hypothesis, found they dropped weak ideas more often, pivoted in a more focused way and earned higher revenue on average; a later multi-trial replication of 759 firms found a similar pattern with fewer pivots. The evidence supports thinking like a scientist, not any specific toolkit.
  One caution is arithmetic. Small samples are noisy, and a founder who sees "3 of 40 people clicked" does not know much. If the true click rate is 10%, a test with 40 visitors can easily show 3% or 18%. Try it: choose a true conversion rate and a sample, and run twenty identical tests.
  
    Interactive · How noisy is a small test?
    Visitors in each test
      
    True conversion rate
      
    Run 20 tests
    
    
    Simulated with random numbers: every press gives different results. The 95% range is the Wilson score interval for the last test.
  
  At 10% and 40 visitors the observed rate across twenty tests typically wanders between about 2% and 18%, and a single test of 4 conversions gives a 95% range from 4% to 23%. At 2,000 visitors the same test pins it within roughly a point and a half either way. The lesson: before believing a result, ask how many people it is based on.

  Unit economics: does one customer pay back?
  A business is a pile of customers, so the question to ask first is about one. The contribution margin is what a customer brings in per month minus the variable cost of serving them. Churn is the share who leave each month, which makes the average customer last 1 ÷ churn months. Multiply and you have the lifetime value (LTV). Against it stands the customer acquisition cost (CAC): all the marketing and sales spend per new customer.
  LTV = contribution margin per month ÷ monthly churn  ·  payback = CAC ÷ contribution marginassumes constant churn and ignores discounting; a rough tool, not a forecast
  A subscription costs $30 a month, $6 to serve, and loses 5% of customers a month. Contribution is $24, the average life 20 months, LTV $480. If acquiring a customer costs $200, LTV is 2.4 times CAC and the customer pays back in 8.3 months. A common rule of thumb in start-up circles asks for an LTV of at least three times CAC and a payback within a year; those are conventions, not laws, and they depend on how long you can wait for cash. Play with it: cut churn and watch the lifetime grow much faster than you expect.
  
    Interactive · One customer's arithmetic
    Price per month
      
    Cost to serve per month
      
    Customers lost per month
      
    Cost to win a customer
      
    
  
  If the price does not even cover the cost to serve, no volume rescues you, which is the break-even lesson in Accounting (34/50). If it covers it but LTV is below CAC, every customer you win loses money, and growth makes the hole bigger faster. This is the check that most business plans skip.

  Runway and burn: how long you have to learn
  Burn is the cash a company spends each month beyond what it earns; runway is how many months the cash lasts at that rate. A company with $250,000, spending $40,000 a month and earning $10,000, burns $30,000 and has about 8 months, if nothing changes. But revenue is rarely flat. Paul Graham's essay Default Alive or Default Dead? (October 2015) gives the sharper test: given current spending and growth, will the company reach profit before the money runs out, with no further fundraising? If yes it is default alive; if no, it is betting on investors.
  
    Interactive · Alive or dead by default?
    Cash in the bank
      
    Monthly spending
      
    Revenue now, per month
      
    Revenue growth per month
      
    
    
    Invented figures in $ thousands. Spending is held flat, which is generous: most teams hire as they grow.
  
  With the starting numbers, cash goes negative in month 11 and dips to about −$32,000 before revenue overtakes spending in month 15. Raise growth to 15% and the company survives with a trough of about $53,000 left. The gap between those two worlds is five points of monthly growth, which is why founders sweat the growth rate and why runway should be counted in months of learning, not dollars.

  Funding paths, and the power law
  Money has a price, and each source charges a different one.
  
    PathYou give upFits
    Bootstrapping (customers fund you)SpeedBusinesses that can earn from day one
    Loans (banks, government-backed schemes such as the US SBA or the UK Start Up Loans)Fixed repayments; often a personal guaranteeSteady, predictable cash flow
    Angels (individuals)Some equity, adviceEarly stage with a story but little revenue
    Venture capitalA large stake, board seats, pressure to grow fastCompanies that could become very large
    Grants and crowdfundingTime spent, public exposureScience, social missions, consumer products
  
  Venture capital is the one that shapes the culture. Investors in a fund expect most of their companies to return little, and a few to return enormous multiples. Peter Thiel argues in Zero to One that in a good fund the best investment returns more than all the others combined; whether or not that holds in every fund, the shape is a power law: a heavy tail where a few outcomes dominate. It explains why VCs push for huge growth and "swing for the fences" even when a solid small business would be a fine life for the founder. The two goals are not the same, and the founder should know which game they are in. As a dated example, Y Combinator's standard deal (as stated on its website in November 2026) is $500,000: $125,000 for 7% of the company on a post-money SAFE plus $375,000 on an uncapped MFN SAFE.
  Here is a toy fund of 20 companies where half return nothing and the rest follow a heavy-tailed distribution. The numbers are invented to show the shape, not to describe real funds.
  
    Interactive · A toy venture fund
    Tail heaviness (lower = heavier)
      
    Raise a new fundSimulate 1,000 funds
    
    
    Illustrative toy: each company is a wipe-out with 50% chance; otherwise it returns at least 0.5× with a Pareto tail capped at 300×. Not data about real funds.
  
  With the default tail, the average fund of 20 returns about 1.4 times the money, but the median fund returns only 0.88 times, and 56% of funds lose money. In a typical fund the single best deal makes up around 40% of everything returned. A few huge winners carry the average, which is how a strategy can look great on average and still lose for most players.

  Equity and dilution: owning less of more
  Selling shares funds the company but shrinks your slice. If an investor puts in $1 million at a $4 million pre-money valuation, the company is worth $5 million post-money and the investor owns 20%; everyone else is diluted by 20%. Add an option pool for future hires, usually set aside before the money arrives so that it dilutes only the existing owners, and a second round, and the founders' share drops again. Founders do this on purpose: 28% of a $25 million company beats 100% of nothing. A SAFE (a simple agreement for future equity, created by Y Combinator) delays the price until a later round but dilutes just the same.
  Investor's share = amount invested ÷ post-money valuationpost-money = pre-money + amount invested. Everyone else owns (1 − investor's share) of what they had.
  
    Interactive · A cap table through two rounds
    Founder A's share of the founders' stake
      
    Option pool after the seed round
      
    Seed round: amount raised
      
    Seed round: pre-money valuation
      
    Series A: amount raised
      
    Series A: pre-money valuation
      
    
    
    $ millions. The option pool is created in the seed round and is counted in its pre-money valuation, so only the founders are diluted by it. Invented terms.
  
  With the defaults, the two founders start with 100% and end with 56% between them (28% each) after the seed and Series A, while the company is worth $25 million after the Series A: $14 million on paper for the founders together. How the founders split the company between themselves matters just as much; Co-Founders: Choosing, Splitting, Staying Together (04/08) covers that.

  Why start-ups fail: the sourced data
  Start with the base rate, because every founder is an exception in their own mind. The US Bureau of Labor Statistics follows every new employer establishment through a cohort. For those opened in March 2013, 50.6% were still operating five years later and 34.7% after ten; about four in five survived the first year. The caveats are real: this counts new workplaces, not venture-backed start-ups; closing is not always failure (the owner may retire or sell); and the figures are for the US.
  The causes are harder to measure. CB Insights collected the post-mortem notes written by founders of failed start-ups. In its 2018 report of 101 such cases, the most common cause, cited by about 42%, was "no market need": building something nobody wanted enough to pay for. Later editions found running out of cash or failing to raise more was a leading cause, in roughly 29% of 242 cases. I could only check these figures through press summaries of the report, not the report itself. They also come from founders explaining their own failures, mostly in venture-backed companies, so treat them as indicative. Taken with the experiments, unit economics and runway above, the pattern is plain: build something people want, make sure each customer pays back, and keep enough months to find out.

  What a graduate just sees
  
    The cheapest way to be wrong is to talk to customers before building anything.
    Compliments are free; deposits, pre-orders and past behaviour are evidence.
    Small tests are noisy. Always ask how many people a result rests on.
    Look at one customer first: contribution margin, payback and lifetime value against the cost to win them.
    Runway is measured in months of learning, and "default alive" is the test of whether the plan survives without a rescue.
    VC money is for businesses that could be enormous; it suits some founders and is wrong for most.
    Every round of funding trades a slice of the company for time. Know the price in dilution.
    Most new businesses do not reach year ten. Survivors mostly learned faster than they spent.
  

  Where it shows up next
  
    Product Innovation & Go-to-Market (34/63) and Pricing (34/61): how the product is shaped, priced and launched.
    Organizational Behavior (34/53) and Business Law & Ethics (34/55): hiring, co-founder disputes, company forms and what founders owe investors.
    Corporate Finance (34/51): valuing a company that earns nothing yet; the discount rates behind a venture investor's demands.
    In the world: a corner shop, a software start-up, a spin-out from a university lab, an internal team inside a large company (see Starting Something Inside a Company, 04/09).
  

  If you want to go deeper
  
    📘 Textbook. Bill Aulet, Disciplined Entrepreneurship: 24 Steps to a Successful Startup (Wiley, 2013). The step-by-step method taught at MIT.
    📘 Textbook. Steve Blank and Bob Dorf, The Startup Owner's Manual (K&S Ranch, 2012). The customer-development process in detail.
    📘 Gentler. Rob Fitzpatrick, The Mom Test (2013), on asking questions that give real answers; Eric Ries, The Lean Startup (Crown, 2011); and Paul Graham's essays at paulgraham.com.
    🌐 Free course. MIT OpenCourseWare 15.390 New Enterprises (Spring 2013, instructors William Aulet, Howard Anderson and Matt Marx): the full course with readings and assignments.
  

  
    The takeaway
    Start with a problem, and look for evidence in what people give up, not what they say. Test cheaply, with enough people to trust the answer. Check that one customer pays back before you pay for a thousand, and count your runway in months of learning. Funding trades ownership for time, and most new ventures fail for reasons visible in the first year.
    Try this: pick one small idea you have. Without building anything, find five people who have the problem and ask each what they did the last time it happened. Write down what they did, not what they said.
  

  
    Sources & Further Reading
    Sources. Lean start-up: Ries coined the term in 2008 and published The Lean Startup in September 2011; critics (Horowitz 2010; Felin, Gambardella, Stern and Zenger) as summarised on the Wikipedia article on the lean start-up. Camuffo, Cordova, Gambardella and Spina, "A Scientific Approach to Entrepreneurial Decision Making: Evidence from a Randomized Control Trial", Management Science 66 (2020), pp. 564–586; replication of 759 firms in four trials, Camuffo et al., CEPR DP15972 (working paper). BLS: "34.7 percent of business establishments born in 2013 were still operating in 2023", The Economics Daily (2024), with the 5-year (50.6%) figure from the same table. CB Insights, "The Top 20 Reasons Startups Fail" (2018 and later updates): figures taken from press summaries (the 42% is from the 101-case 2018 report; 29% is from a later edition's 242 post-mortems), the original page not directly checked. Paul Graham, "Default Alive or Default Dead?", October 2015 (read on paulgraham.com). Y Combinator's deal: ycombinator.com/deal, read in November 2026. Thiel's claim is from Zero to One (2014) and is quoted as his argument, not as measured fact. MIT 15.390 page confirmed (Spring 2013). The Mom Test is cited for its ideas; edition details not verified. All interactive numbers were computed in node: Wilson 95% interval for 4 of 40 = 4.0% to 23.1%; unit example LTV $480, ratio 2.4, payback 8.33 months; runway default cash < 0 in month 11 (low −$32k), at 15% growth trough +$53k; toy fund over seeds 1 to 1,000: mean 1.44×, median 0.88×, 56.5% below 1×, best deal about 40% of the total; cap table post-seed 5.0, founders 70%, post-A 25.0, founders 56%, $14.0 million.
    
      🌐 MIT OCW 15.390 New Enterprises
      🌐 BLS: business establishments born in 2013
      🌐 Paul Graham: Default Alive or Default Dead?
      📘 Aulet, Disciplined Entrepreneurship