ROI of a pricing solution: 

Time savings or profit margins?

Two types of ROI that are fundamentally different

When a company applies AI to a function, it reaps two very different benefits, and it tends to lump them together in the same table.

The first is the time saved. Hours that used to be spent consolidating a database, following up with a data provider, or organizing a committee—and that are no longer spent on those tasks. Let’s call this “friction ROI.” It’s calculated in hours per week multiplied by an all-in cost, is evident from the very first month, and immediately resonates with an executive committee.

The second is a better decision. A price set two cents higher on a product where the customer won’t notice it, a markdown triggered three weeks earlier, a promotion that didn’t cannibalize the product next to it. Let’s call this the ROI of decision-making. It’s measured in margin points, takes one or two quarters to become clearly visible, and is much harder to attribute.

These two ROIs do not have the same properties, and that is where it all comes down to.

The ROI from friction is capped, but the ROI from decision-making is not

A four-person pricing team cannot generate more than the equivalent of four full-time employees. That’s an arithmetic limit. Even if you automate all administrative tasks, the reduction in friction stops where the team’s payroll ends.

Margin gain, on the other hand, is calculated based on revenue. For a company with 400 million euros in revenue, a half-point increase in margin represents two million euros. No automation of administrative tasks will ever reach that magnitude. The two ROIs simply aren’t in the same league.

Friction-based ROI is certain; decision-based ROI is probabilistic

And yet, the first one is almost always the only one included in the business case, because it’s the only one we know how to write without getting nervous. Six hours per week multiplied by fifty-two weeks multiplied by an hourly rate: the figure is clear, defensible, and no one will challenge it in a committee meeting. It’s also the only one of the two that appears in most maturity models, including the one described in our guide, “Automate Your Pricing Without Losing Control.”

Margin gains, on the other hand, depend on assumptions about elasticity, scope, and duration, and are subject to a myriad of external factors. As a result, they’re either written in fine print—or not mentioned at all. The outcome is predictable: a project is signed off based on hours, evaluated eighteen months later based on margin figures, and no one ends up getting what they expected.

This figure is worth considering. It doesn't mean that AI is useless. It means that two-thirds of organizations do see an impact, but only a minority are able to translate that into a return on investment. Between the two, there's a conversion process that doesn't happen on its own.

Friction ROI: Easy to Measure, Easy to Overestimate

Friction ROI is a valid concept. In a pricing team, administrative tasks take up a significant portion of the week—and we’ve devoted an entire article to the actual breakdown of a pricing team’s time. The problem isn’t counting it. The problem is how we count it.

Three Common Mistakes in Calculating Hours

• Count the hours eliminated rather than the hours reallocated. A task that is eliminated only creates value if someone uses that time for something else. If the six hours freed up are taken up by other meetings, the gain is zero in practice, even if it appears to be real on paper.
• Forget about the time the tool takes. Configuring, monitoring, resolving edge cases, training new hires: a pricing system creates its own workload. The net result is always positive with a successful implementation, but it’s never equal to the gross.
• Apply the executive's hourly rate to managerial tasks. That's the classic trick that inflates a business case by 40% without anyone noticing.

What Saving Time Really Brings

Its true value isn't financial—it's strategic: it changes what the team is allowed to do. A team that spends its week producing files doesn’t have time to test, simulate, or challenge a rule. A team freed from administrative tasks can finally focus on pricing. The time saved is therefore not an ROI in and of itself; it’s the prerequisite for achieving the second ROI.

To put it another way: time saved is a means to an end; profit gained is the end goal. A business case that stops at the means stops halfway.

Decision ROI: Hard to Measure, Yet the Only One That Really Matters

The increase in margin stems from four categories of decisions, and it is helpful to distinguish between them because they are not measured at the same rate or within the same scope.

1. The best-selling shelf-stock items

For products that customers don’t compare, setting a price that’s too low means giving up profit margin without gaining any brand equity in return. This is the largest and slowest-growing source of profit: it’s earned in small increments across thousands of products, and it requires the ability to distinguish between what customers look at and what they don’t.

2. Less Destructive Promotions

A promotion that generates high sales isn't necessarily a profitable one. Profit comes from the ability to anticipate cannibalization and measure actual incremental sales rather than apparent volume.

3. A price reduction at just the right time

A three-week lead heading into the end of the season is often worth more than five extra points at the very end. It’s the quickest gain to see and the easiest to isolate, making it an excellent area for a driver to focus on.

4. Errors That No Longer Occur

A price entered with a misplaced decimal point, a rounding rule that pushes the price below the break-even point, a promotion that remained active three weeks after the campaign ended. This kind of profit isn’t visible in the spreadsheets, since it relates to events that never happened. Yet it’s often the first to materialize.

How to Measure a Margin Increase Without Beating Around the Bush

This is where most projects go off the rails, and it can be summed up in one sentence:comparing the before and after proves nothing.

Between the two periods, there was a season, weather conditions, inflation, a competitor that opened or closed, a change in search rankings, and a marketing campaign. Attributing the difference to the tool is a matter of belief, not measurement. And since the discrepancy can just as easily work against the tool, this method produces just as many false failures as it does false successes.

The only method that works: the control group

You need two comparable groups—one under the new system and the other remaining under the previous system—over the same period. The difference between the two is the only figure that can be attributed to the change.

• Create groups that are truly comparable. Same store type, same competitive area, same seasonality. Two stores of the same size in two different catchment areas do not constitute a control group.
• Maintain the duration. At least one full cycle for seasonal categories, and one quarter for core inventory. Anything less than that is just noise.
• Finalize the protocol before starting. The scope, metrics, and evaluation date must be established before the project begins. Choosing metrics after the fact guarantees a positive result—but one that no one will believe.
• Measure margin and price perception together. A margin gain achieved by damaging price perception is a liability, not a gain. Both metrics are tracked on the same dashboard, which we detail in “Pricing KPIs: Building Your Retail Dashboard.”

A realistic estimate

With a properly executed rollout, the expected gain ranges from +0.5 to +3 margin points within a few months, depending on the starting point, the quality of the benchmark, and the retailer’s actual pricing flexibility. A promise that starts above this range warrants a request to see the measurement protocol. A promise that provides no range at all warrants the same question.

Building a business case that still holds up twelve months later

A useful pricing business case separates the two ROIs, adds them together at the end, and assigns a different confidence metric to each one.

The Crossed-Out Line Test

Take the business case you were given and cross out all the lines that relate to time saved. Look at what's left.

If it remains a profitable project, you’re buying a decision-making tool, and the time saved will be a bonus. If there’s nothing left, you’re buying a productivity tool: that’s perfectly legitimate, but you should then compare it to other productivity tools, not to a pricing solution, and the budget is different.

The Three Lines Nobody Writes

• The Cost of Data. Ensuring the reliability of a product database, maintaining competitive matching, and integrating data sources: this is an ongoing effort, and it is the true prerequisite—as our article on data quality points out —for breaking through the glass ceiling of AI pricing.
• Maintenance costs. A pricing system is not a static asset: its set of rules becomes increasingly complex, contradictory, and ultimately costly if no one reviews it. This is the subject of our article on the “rule debt” of a pricing engine.
• The exit cost. The amount you’ll receive if you change your mind in three years is negotiated at the time of signing—never afterward—and is factored into the initial business case. See the reversibility of a pricing solution.

These three steps take place during the scoping phase, not at the time of signing: this is the first phase described in our guide to implementing a pricing tool.

The Right Reading Perspective

Friction ROI becomes apparent after three months. Decision ROI becomes apparent after two to four quarters, because you have to let a full cycle run its course. Announcing both at the same time sets the stage for disappointment at the first monitoring committee meeting: the hours will have been put in, but the margin won’t be clear yet, and confidence in the project will hinge on that very moment.

So the process that works is always the same: announce the hours three months in advance, announce the margin twelve months in advance, and make this clear from day one.

FAQ