The 5 critical mistakes of a Revenue Management System (RMS)

In brief

An RMS should never mix budget and pricing, mechanically copy the competition, shift the responsibility for building its own rules onto you, learn from individual user corrections, or disappear once the contract has been signed. These five mistakes reveal structural limitations of the product, not configuration choices.

The RMS market has reached maturity. Demos are becoming increasingly impressive, interfaces increasingly modern and marketing promises increasingly ambitious.

Yet, when we support hotel groups in choosing an RMS, we regularly come across the same warning signs. Elements that may seem appealing during a demo but that, in our view, reveal structural limitations of the product.

Here are the five criteria that should, at a minimum, prompt you to ask yourself some serious questions before choosing an RMS.

Criterion 1: Your RMS takes the budget into account when recommending a price

This is probably the most important conceptual mistake.

If an RMS displays an indicator such as “Price needed to reach budget”, it mixes two concepts that should never be combined: financial management and the determination of the price.

The budget is a management objective. It is essential for managing a hotel business, but it creates absolutely no demand. No guest becomes more willing to pay because a hotelier set a revenue target for a given date 18 months earlier. Price should be the consequence of demand, perceived value and market context. Never of a budget target.

Concretely:

A hotel has four rooms remaining and wants to generate €1,000 in revenue. The RMS tells them they should sell at €250.

• Why stop there?

Simply change the budget to €1,600 and the same reasoning will lead to a recommendation of… €400.

• The absurdity is immediately apparent.

At Revbell, we have deliberately chosen never to display this indicator. Because it encourages bad practices rather than helping people make better decisions.

The budget remains a compass. It should never become a pricing engine.

Criterion 2: Your RMS promotes automatic alignment with the competition

“Automatically position yourself €1 below your competitor.”

This feature is still widely promoted by some vendors.

Yet, if it is one of the product’s main selling points, it is often a sign that the analytical engine is unable to produce an independent recommendation.

Mechanically following your competitors has never been a Revenue Management strategy.

• An RMS must understand the market.

It must be able to identify the moments when you need to be cheaper than the competition… but also those when you need to be significantly more expensive.

The goal is not to follow the market. It is to optimize your own value.

At the extreme, if your only strategy is to systematically stay €1 cheaper than your competitors, a third-year intern with no RM expertise, equipped with a rate shopping tool and five minutes every morning, could probably achieve a similar result.

That is not what we expect from an artificial intelligence software.

Criterion 3: The main added value of the RMS is creating your own business rules

This is often presented as an advantage.

“You can create as many rules as you want.”

We see it more as a warning sign. Of course, some business rules remain useful for translating certain operational constraints. But when an RMS primarily promotes its rules and workflow engine, it is shifting part of its work onto you.

Those who remember the large Excel files of the 2000s know exactly how this ends:

• You add a rule.
• Then an exception.
• Then an exception to the exception.

A few years later, no one really understands why the system produces a particular recommendation. An RMS is not supposed to be an advanced macro builder. It is supposed to solve complex problems through its analytical intelligence.

Ultimately, the question is simple:

If I am the one building the rules that create the value, shouldn’t I be the one charging the RMS?

Criterion 4 : Recommendations learn from user corrections

This is a question that comes up very often during RFPs.

“Does your AI learn from our validations or corrections?”

Our answer is always the same: Certainly not. Why?

Because the purpose of an RMS is not to reproduce human decisions. Its purpose is to outperform them.

If every change made by a Revenue Manager were to modify the recommendation algorithm, the system would gradually incorporate biases, habits and sometimes… mistakes.

Do we really want the least experienced member of the team to permanently influence the intelligence of the engine?

Obviously not.

On the other hand, some forms of learning are perfectly legitimate.

• The forecast can learn from changes in observed behavior.
• Forecasting models can be continuously recalibrated.
• The parameters of AI models can be adjusted using aggregated data and observed performance.

But pricing recommendations should never learn from individual user corrections.

Otherwise, the RMS gradually stops being a decision-support system.

It simply becomes a reflection of your habits.

Criterion 5: The vendor neglects Account Management

The best algorithm in the world will not create value if it is poorly implemented.

Choosing an RMS is therefore never just about the software.

You must also assess the quality of the teams that will support you over several years. And ask very specific questions.

• What is the profile of the Account Managers?
• Do they have genuine Revenue Management experience?
• Which languages do they speak?
• How does the implementation process work?
• Who supports you during the initial configuration?
• How often are training sessions organized?
• Is there follow-up several months after deployment?
• Do the support teams genuinely understand the product’s algorithmic mechanisms?

These questions are sometimes more important than some of the features highlighted during a demo.

In hospitality, an RMS is never just a piece of software. It is a performance partner.

And like any partner, the quality of the human relationship matters just as much as the quality of the technology.

Conclusion

• An excellent RMS does not ask you to compensate for its limitations.
• It does not ask you to build its rules.
• It does not ask you to blindly follow your competitors.
• It does not ask you to constantly correct its recommendations.
• And it does not disappear once the contract has been signed.

On the contrary, a good RMS knows how to analyze demand, understand your market, produce robust recommendations and support you sustainably in their implementation. That is the best test to apply before signing.