One Price Leaves Money on the Table: The Economics of Segmented Pricing

Every product has customers who would happily pay more than its price and customers who would buy it if it were a little cheaper. A single price serves neither. It charges the enthusiast less than the product is worth to them and turns away the price-sensitive buyer whose business would still be profitable. Microeconomics has a name for the money lost on both sides, and a set of tools to recover it.

The rectangle under the demand curve

Picture demand as a line: at 100 nobody buys, and each unit of price lower adds a thousand customers. The product costs 20 to make. With one price, the profit-maximising choice is 60: 40,000 units at a 40 margin, a profit of 1.6 million. That profit is a rectangle under the demand curve. The triangles around it are value the company leaves behind: willingness to pay above 60 that it does not charge, and customers between 20 and 60 that it does not serve.

Offer three versions instead, at 80, 60 and 40, and suppose each customer buys the version that matches their willingness to pay. The company sells 20,000 units at each price and earns 1.2 million, 0.8 million and 0.4 million: 2.4 million in total, 50% more than with one price, from the same demand and the same cost. The steps fill the space under the curve that the single rectangle left empty.

Three degrees of price discrimination

Economists, following Arthur Pigou, distinguish three ways of charging different prices for essentially the same product:

• First degree: each customer pays exactly what it is worth to them. It is the theoretical limit, approached in negotiated B2B deals and auctions.
• Second degree: the company offers a menu and customers choose: versions, bundles, volume discounts, good-better-best tiers. It needs no information about the individual customer.
• Third degree: different prices for observable groups: students, regions, channels, business versus consumer, time of purchase.

The best-documented corporate case is airline yield management. In the 1980s American Airlines built a system to sell the same seat at different prices depending on when and how it was booked, with restrictions that kept business travellers from buying the cheapest fares. The company later estimated its quantifiable benefits at about 1.4 billion dollars over three years.

Fences make it work

Segmented pricing fails when the customers who would pay 80 can simply buy the version at 40. The design problem is the fence: a difference that matters to one segment and not to the other.

1. Product fences: features, capacity, service level or support that the high-value segment needs and the price-sensitive one does not.
2. Time fences: early booking, off-peak use, delayed availability.
3. Channel and volume fences: prices that depend on how and how much the customer buys, which also tend to reflect real differences in cost-to-serve.
4. Customer fences: eligibility rules for groups that are verifiable and defensible, within the limits of competition and consumer law.