Most revenue managers don't know why their strategy worked.

congratulations, you're a genius!

Imagine you're playing poker with your friends. You've had a few drinks, you're holding a questionable hand, and for reasons even you don't fully understand, you decide to go all in.

You win.

Fantastic. Clearly you're a poker genius. You've mastered probability, human psychology and risk management. You might even consider quitting your job and moving to Las Vegas. Except you didn't actually make a particularly intelligent decision. You got lucky.

Now imagine you lose the next hand despite having made a statistically excellent decision. Does that suddenly make you an idiot? Of course not.

Yet somehow, in revenue management, we have become incredibly comfortable judging the quality of our decisions almost entirely by their outcomes. And when the outcome is good, we take the credit.

the market grew. apparently, so did your intelligence.

Your hotel finishes the month with ADR up 12%, RevPAR up 15% and market share improving. The revenue meeting is a beautiful celebration of strategic excellence. Our pricing strategy worked. Our segmentation was brilliant. Our distribution decisions were spot on. The RMS recommendations delivered exceptional results. Wonderful.

But what if the market grew 20%? What if your biggest competitor had 100 rooms out of order? What if a major event generated demand nobody anticipated? What if an OTA algorithm suddenly improved your visibility?

Or perhaps your strategy actually left money on the table, and you would have achieved 18% RevPAR growth by doing absolutely nothing. Would you still call it a success?

That's the problem with looking at results without understanding what actually caused them.

A good result doesn't prove a good strategy. It only proves you achieved a good result.

we're brilliant at explaining things afterwards

Human beings have a fascinating ability to create explanations once we already know what happened.

If occupancy increased after we reduced rates, the discount obviously worked. If occupancy increased after we raised rates, our premium positioning clearly worked. If we held rates and occupancy increased, our patience was rewarded. Notice something?

Whatever happened, we can usually find a story that makes us look intelligent. This is called hindsight bias, and revenue management is absolutely full of it.

We sit in meetings reviewing historical performance, connecting dots and creating beautifully logical explanations for decisions that were made under considerably more uncertainty.

The danger is that we start believing those explanations. And once we believe them, we repeat the strategy. Because it worked last time......Apparently.

correlation is a dangerous little minx

Let's say you increased your rates by 10% and subsequently achieved higher revenue.

Was the rate increase responsible? Possibly.

But perhaps demand was already strengthening. Perhaps your competitors increased their rates even more. Perhaps your channel mix shifted towards higher-paying customers. Perhaps your previous rates were already too low.

Now reverse the example: You launch a promotion, bookings increase by 30%, and everybody celebrates the campaign.

But how many of those customers would have booked anyway? How many existing customers simply paid less? How many bookings were displaced from another channel? And what happened to profitability after acquisition costs?

Without asking those questions, we're measuring activity and outcomes while making assumptions about the relationship between them.

AI will make us even better at being wrong

This is where things get particularly entertaining. AI can now analyse enormous amounts of data, identify patterns and generate explanations in seconds.

Ask it why your hotel performed well, and it can produce a wonderfully convincing analysis covering pricing, demand, segmentation, booking windows, competitor behavior and market conditions. It might even include a lovely graph. But a convincing explanation isn't necessarily a correct explanation.

If the underlying data doesn't allow us to isolate the impact of a decision, AI cannot magically manufacture the missing counterfactual. It can help us test hypotheses, identify competing explanations and estimate possible effects, but it should also be able to tell us when the evidence isn't sufficient.

Unfortunately, humans love certainty. And AI is incredibly convenient at providing answers that sound certain. Which means we're potentially entering a world where revenue managers who never understood why their strategy worked now have an AI agent confidently explaining why it worked.

Wonderful. We've automated the bullshit.

so how do we actually know?

We probably won't know with absolute certainty every time. Revenue management operates in a constantly changing environment, and we rarely have the luxury of perfectly controlled experiments. But we can become considerably better at understanding our decisions.

Before changing a price, launching a promotion or adjusting distribution, document what you believe will happen and why. Define what you're trying to influence, what other factors might affect the result and what evidence would suggest your hypothesis was wrong.

Then measure what happened. Not just whether revenue increased, but whether the behaviour you intended to influence actually changed.

Did the promotion create incremental demand? Did the rate increase genuinely improve profitability? Did the distribution change improve your customer mix, or did it simply move bookings between channels?

And sometimes, after all that analysis, the correct conclusion will still be: We don't know.

That's perfectly acceptable. It's considerably more intelligent than inventing certainty.

maybe we should stop celebrating ourselves so quickly

I sometimes wonder how many revenue managers are genuinely good at revenue management and how many have simply been fortunate enough to operate in strong markets.

Equally, how many excellent revenue managers are being criticized because they're operating in markets where demand is declining despite making commercially sound decisions?

We love rewarding outcomes because they're easy to measure. Understanding decision quality requires considerably more effort. But if we want revenue management to evolve into genuine commercial intelligence, we need to get comfortable separating skill from luck, strategy from market conditions, and correlation from causation.

Because the most dangerous revenue manager isn't necessarily the one who makes a bad decision. It's the one who makes a bad decision, gets a good result, and then spends the next five years repeating it because they're convinced they're a genius.

Winning doesn't automatically mean you played well. And losing doesn't automatically mean you played badly.

The interesting question is whether you understand the difference.

Love, Fabi