Yield management outside airlines: what transfers and what doesn't

Yield management grew up in airlines for a reason. A seat on a flight is perishable, capacity is fixed in the short term, and customers book at different times with very different willingness to pay. The classic account is Smith, Leimkuhler and Darrow's Yield Management at American Airlines, which describes breaking the problem into three parts: overbooking, discount allocation and traffic management.

Plenty of businesses outside aviation now say they "do yield". Some genuinely benefit. Others have borrowed the vocabulary and the booking curves without the conditions that make them work. Before you build anything, it is worth separating the parts of yield thinking that travel from the parts that only make sense when you have a plane to fill.

What transfers almost everywhere

Segmentation with fences. The core idea isn't the algorithm. It is that different customers value the same product differently, and that you can offer different prices if a credible fence separates them: advance purchase, flexibility, size of transaction, channel, contract length. Without a fence, a lower price simply leaks to customers who would have paid more.

Measuring realised price, not list price. Revenue managers judge themselves on what was actually earned per unit of capacity. That discipline transfers directly. Anyone setting rates should know what the business really took home after discounts, overrides and channel costs.

Forecasting before pricing. A good yield system forecasts demand first, then prices against the forecast. Even where there is no capacity constraint, a forecast of volume by segment tells you what a price move is likely to do.

Governance. Floors, change limits, override logs and a weekly review work the same way whether the price is a fare, an FX spread or a seat licence.

What only transfers where capacity is fixed and perishable

Booking curves and inventory controls. Protecting seats for late, high-value demand only makes sense if unsold capacity disappears. Hotels, tours, car hire, events and freight share that property. Many other businesses don't.

Overbooking. It depends on no-shows and fixed capacity. Outside those conditions it is a solution to a problem you don't have.

Displacement cost. In airlines, selling a cheap seat today may displace a full-fare passenger later, so the true cost of a sale includes that opportunity cost. Where supply is effectively unlimited, displacement is close to zero, and pricing should focus on margin, segmentation and leakage instead.

For the full toolkit and its assumptions, Talluri and van Ryzin's The Theory and Practice of Revenue Management is still the standard reference. It is useful partly because it is explicit about which models need capacity constraints and which don't.

Applying it: three industries

FX

Retail and SME foreign exchange has no seat to fill, but it does have something yield-like. Prices are spreads over a moving reference rate, and they differ by channel (branch, online, partner, airport), by transaction size and by customer type. The wholesale market behind the reference rate is enormous: the BIS Triennial Central Bank Survey of foreign exchange turnover put global FX trading at $9.6 trillion a day in April 2025. The reference rate isn't the margin problem. The spread structure on top of it is.

What transfers: segmentation by size and channel, with fences customers understand; rate-setting rules that move with the reference rate inside guardrails; competitor rate checks for the channels where customers compare. Where there is physical currency, inventory matters too. Stock has a holding cost and an ordering cost, so supply chain thinking sits alongside pricing.

What doesn't: booking curves. A customer buying currency next week isn't consuming perishable capacity.

Travel

This is yield management's home ground, and most of it transfers directly to hotels, tours and other capacity businesses. The common gaps are elsewhere: ancillary pricing that isn't managed with the same rigour as the core fare or room, distribution costs not netted off when comparing channels, and too many manual overrides that nobody reviews.

Illustrative example. A tour operator runs a capable revenue management system for departures, but prices optional extras from a static sheet updated once a year. The yield logic is strong where the system is and absent where it isn't. The first gains often come from the parts the system never touched.

Subscriptions

Software and subscription businesses have near-zero marginal cost and no perishable inventory, so classic capacity yield doesn't apply. What does transfer is the fence logic. Tiered packaging, of the kind Rafi Mohammed sets out in HBR's The Good-Better-Best Approach to Pricing, is a way of letting customers sort themselves by willingness to pay without deep discounting.

What doesn't transfer is the habit of charging loyal customers more because they are less likely to shop around. Regulators have shown how they view that. In UK home and motor insurance, for example, the FCA's rules on setting renewal prices mean a renewal price must not be higher than the equivalent new business price. Even where no such rule applies, the commercial lesson is the same: renewal is where subscription margin is won or lost, and it should be priced on value delivered, not on inertia.

Yield thinking pays where it fits the economics of the business. Where it doesn't, the margin is usually in segmentation, leakage and governance rather than in a more complicated model.

Examples marked illustrative are hypothetical. Third-party figures are from the sources linked.