Revenue Went Up. That Tells You Almost Nothing.

Ask a commercial team how revenue management is performing, and the answer is usually a revenue figure. Takings are up four percent on last year, so the function must be working. It is the most natural inference in the world, and in my experience it is wrong about as often as it is right.

Revenue is an outcome. It gets shaped by demand, capacity, the wider economy, timetable changes, what competitors chose to do that quarter, even the weather. Most of those forces have nothing to do with the quality of the pricing decisions a team actually made. Judging revenue management by the revenue line is a bit like judging a navigator by whether the tide came in. Sometimes the two move together. That does not make one the cause of the other.

If you want to know whether your revenue management is genuinely working, you have to measure the thing you control, which is the quality of the yield decision, and hold it apart from the things you do not. This is harder than reading the top line. That is exactly why so few teams do it. But measuring properly is what turns revenue management from a reporting function into a managed one.

Decompose before you celebrate

Start by breaking any revenue change into its two real drivers: how many people travelled, and how much each of them paid. Volume and yield.

A four percent revenue rise built on five percent more passengers paying one percent less per journey is a completely different story from the same rise built on flat volumes and four percent better yield. In the first case demand grew and your pricing actually slipped. In the second, the pricing got sharper. The headline number is identical. What sits underneath it is the opposite.

Yield versus volume decomposition is the foundational diagnostic of the discipline, and it belongs at the top of every performance review, before anyone reaches for another metric. It changes the conversation, because it forces a team to attribute the change rather than just observe it. And once revenue is split this way, the obvious next question almost asks itself: was the yield movement something we did, or something the market did to us?

Strip out what you did not control

A fair measure of performance has to hold the external factors constant. Demand is the big one. If the market grew, flat yield is not a neutral result; it can be a failure to capture conditions that were handed to you on a plate. If the market softened, a small yield decline might actually be excellent defensive pricing against a falling tide. The same yield number means different things depending on the demand it was earned in, and any honest framework has to account for that.

The same goes for capacity changes, timetable revisions, and competitor moves. When a rival pulls capacity off a corridor, yields on that corridor will climb no matter how well or badly you priced. Bank that windfall as revenue management skill and you flatter the function. Worse, you bury the places where the pricing was genuinely poor but got masked by kind conditions. The point of measurement is not a comfortable number. It is to isolate what the pricing decisions themselves contributed, so the good calls can be repeated and the bad ones caught.

The traps that flatter and mislead

Three traps come up again and again.

The first is load factor worship: treating a full train as self-evidently good. A train that fills easily may simply be a train you priced too cheaply. Occupancy and yield pull against each other by design, and a metric that rewards one while ignoring the other will quietly nudge a team into giving revenue away. Watch load factor as an input. Do not manage to it as a goal.

The second is the blended average. Network-level average yield can sit perfectly still while large movements cancel each other out underneath it, strong flows weakening as weak flows improve, netting to nothing you can see. Revenue management is won and lost flow by flow, bucket by bucket. Measurement that only works at the aggregate will never see the decisions that matter. If your analysis cannot get down to the flow, it cannot diagnose anything.

The third is having no counterfactual. "Revenue grew" means nothing until you can answer "compared with what?" And the honest comparison is not last year, because last year was a different world. It is what revenue would have been under the demand and capacity you actually faced, had pricing been left at a sensible baseline. That counterfactual is the only fair benchmark for the value the function added. Build one, even a rough one, and the whole quality of the conversation shifts.

Measure the decision, not the weather

None of this needs a new system or a bigger team. It needs a willingness to ask a harder question. Not "did revenue go up?" but "were the yield decisions good, given everything that was true about the market at the time?"

So: decompose revenue into volume and yield. Normalise for the demand and capacity you were handed. Refuse the metrics that flatter you. Build a counterfactual, however rough, so contribution can be told apart from circumstance.

A team that measures this way will sometimes have to report that a rising revenue line hid weak pricing, or that a disappointing one hid an excellent defensive performance. Those are awkward things to put in front of a sponsor. They are also the only findings worth having, because they describe the part of the result the team can actually control, and therefore the part it can actually improve.