Predictive or reactive pricing: the webinar summary
March 31, 2026
The Barbotteau Group previously used a single " Markup " in Excel. It now calculates three optimal prices per product—based on volume, revenue, or margin—selected according to each category’s role, and achieved a 1-point increase in the " Markup " rating in one quarter for the products tested, with a significant improvement in its price image.
The group made the transition directly from Excel to a hybrid AI system, skipping the intermediate step of using rules. Support from management and a pilot test—which was first limited, then scaled up—helped secure buy-in, while the teams retain control over every decision.
On April 9, 2026, the Pricing Club hosted a webinar focused on a simple question: Is your pricing proactive or reactive? During the webinar, the Barbotteau Group explained how it transitioned from a one-time “ Markup ” in Excel to simulating prices before publication.
Here is what was discussed: the starting point, the method, the test, the results, and the lessons learned about adoption.
Predictive vs. Reactive Pricing: What the Webinar Revealed
Predictive pricing involves simulating the effect of a price on volume, revenue, and margin before setting it. Reactive pricing makes adjustments after the fact, once sales have been recorded.
The webinar on April 9, 2026, did not simply contrast the two in theory. It used a real-world example to demonstrate how switching from one to the other affects a pricing team.
The session, organized by the Pricing Club, was chaired by Victoria Bayen, director of pricing at Conforama.
Mathieu Guimber, director of pricing at the Barbotteau Group, discussed his transformation; Ludovic Shum, a pricing consultant at Booper, discussed the method and the tool.
The remarks, delivered in English, have been translated and condensed here.
The starting point: a single " Markup " in Excel
The Barbotteau Group, based in the Caribbean and French Guiana, oversees pricing for four retail chains, including a supermarket, a garden center, and a local organic store.
For years, the rule was simple: a target " Markup ," applied to almost all products.
The method worked for a long time. It is no longer enough when purchasing power is under strain.
The Guadeloupe market has fewer than 400,000 residents, three major players, and largely overlapping suppliers. The group sells its products there exclusively through brick-and-mortar stores. Every customer knows everyone else’s prices, and the high cost of living weighs heavily on every decision.
Facing a competitor with a pricing team of about ten people, Mathieu Guimber was running the show all by himself.
There is an additional constraint that is rare in mainland France: most products arrive by ship, and a supply disruption can last two months. The right price must therefore also balance the rate of sales—neither too fast nor too slow.
Two areas of focus: price positioning and product assortment
First project: a converted supermarket operating under a new brand name, which had inherited a price image considered disastrous. The goal was twofold: to restore a positive price image and maintain a sustainable profit margin.
Second project: a garden center with an overly broad product selection. An analysis of product categorization revealed that a small number of items accounted for a large share of sales and profit margins.
The goal—which is still in progress—is to reduce the number of SKUs from 20,000 to 12,000 and free up 200 k€ in working capital.
Ludovic Shum explained the reasoning: before discontinuing a product, give it one last chance—a price reduction—and see how it affects sales.
From Forecasting to Decision-Making: Three Optimal Prices
A forecast is useless if it does not lead to a decision. The method presented calculates three prices for each product: the price that maximizes volume, the price that maximizes revenue, and the price that maximizes margin.
The choice between the three depends on the category's role in the store.
The volume-based prize goes to the high-traffic categories—essentials such as pasta and rice. The revenue-based prize goes to the destination categories, which are the store's mainstay.
The markup applies to service products, which are purchased based on need rather than on price comparison.
Within each product family, the products are rated from A to E.
Products in Category A, which are the most visible, are positioned as value-priced items: a tighter margin is accepted for them. The subsequent categories, which are less prominent in the customer’s mind, make up for the margin difference.
Mathieu Guimber emphasized one point: isolating a product and testing its response to an increase or decrease takes him just a few minutes, whereas Excel made the task impractical.
Start small, test, measure, then scale up
The group didn't make the switch all at once. It selected a range of products in the supermarket, tracked the results over one to three months, and then expanded the initiative to the entire supermarket and all four store formats.
“Test it on a small scale, show that it works, and then roll it out: that’s less risky than launching everything everywhere,” he summarized.
Markup s on the products tested over the course of a quarter at the Barbotteau Group’s supermarket, showing a marked improvement in price image.
(results reported by the Barbotteau Group during the webinar; replay from the Pricing Club, April 2026)
On the customer side, complaints about prices being too high—which used to be a recurring issue—have disappeared. On the market side, no competitors had been raising the supermarket’s prices; for about six months now, they have been doing so.
The message that comes across is not “the lowest price” but “the right price.”
Two factors played a key role in the adoption of the system: the explicit support of management, which drove the decision to choose an AI tool, and the test results. In response to the objection, “We’ve been doing it this way for a hundred years, and it works,” the evidence proved convincing.
He also says it straight out: the tool is powerful and takes some getting used to, but the results speak for themselves.
Excel, rules, hybrid AI: You can skip a step
Ludovic Shum described this journey as following the most common path in retail: Excel, then a rule-based tool (if this condition is met, then that price), and finally a hybrid AI.
The Barbotteau Group made the transition directly from Excel to hybrid AI, without going through any intermediate steps based on rules.
Hybrid, in the sense that AI generates proposals while the teams retain control over the rules, strategy, and approval. During the session, no one advocated for full automation.
In-session surveys gauge participants' starting points. According to the results presented live, half of the respondents managed their pricing using Excel, and 40% used an in-house solution.
When asked about their top priority, 30% cited optimizing margins and volumes, 25% cited pricing governance, and 25% cited the speed of price monitoring. The number of respondents was not disclosed.
The Chairperson's Perspective
Victoria Bayen brought her experience with a furniture retailer to the table as it competed against Asian platforms, which had an advantage in terms of purchasing costs.
His response: avoid competing on the same products, rework the sourcing strategy, or offer higher-value products, while maintaining the image of a discount retailer.
For her, Excel is often still the first step: it’s through Excel that a pricing team proves its value to management before investing in a more powerful tool.
According to her, the most difficult part of a transformation is engaging with the procurement and supply teams, who used to set prices and know their products well. Providing insights without making them feel disempowered requires tact.
His conclusion: forecasting and elasticity are powerful tools—but they are just tools. The real work lies in making them understandable, usable, and trustworthy for the teams that set prices every day.
Ludovic Shum, a pricing consultant at Booper, described how the models work: AI—using statistics and machine learning—estimates what might happen in the short and medium term.
Operations research then selects the most relevant scenario.
The forecasts cross-reference each store's sales and prices with external factors: temperature, weather, holidays, and vacation periods. Each factor is assigned a weight, store by store, day by day.
When asked about marketplaces, he explained GENIUS Link’s product matching process in detail: the barcode, when available; otherwise, the product labels; then the image; with human validation to refine the suggestions. See product matching.
Watch the webinar again
The full replay, in English, is available on the Club du Pricing channel.
To learn more about this method, see our guide to AI-powered sales forecasting in retail, and the Barbotteau Group case study in our article on the benefits of a pricing platform for mass retailers.
FAQ
It involves simulating the effect of a price on volume, revenue, and margin before setting it, based on historical sales data and external factors. Reactive pricing, on the other hand, adjusts prices after sales figures are known.
The reactive approach is based on what has already happened: a target margin, alignment with a competitor, or a correction when sales decline. The predictive approach makes decisions in advance by comparing several pricing scenarios and their expected effects.
For each product, the price that maximizes volume, the price that maximizes revenue, and the price that maximizes margin. The choice depends on the category’s role: traffic, destination, or service.
Among the supermarket products tested, customer satisfaction scores rose by 1 point Markup in a single quarter; comments that the products were “too expensive” disappeared; and competitors have since begun raising their prices. The approach was then expanded to all four of the group’s retail chains.
This is the most common path, but it's not mandatory. The Barbotteau Group made the transition directly from Excel to a hybrid AI system, relying on management support and a limited pilot that demonstrated its effectiveness.
The replay, in English, is available on the Club du Pricing YouTube channel:“Is Your Pricing Predictive or Still Reactive?”, published on April 13, 2026.
Sources: Pricing Club webinar on April 9, 2026; replay published on April 13, 2026; remarks made in English, translated and condensed by Booper. The results are those reported by the Barbotteau Group and have been approved for publication. Updated: October 2026.
Booper will be participating in Tech For Retail 2026, the European retail trade show, at Paris Expo Porte de Versailles (Hall 7.2), from November 30 to December 1. The team looks forward to seeing you at booth N78.
On the agenda for this edition: agent-based AI and its real-world impact on headquarters, and unified commerce. At the booth, live price simulations and a competitive intelligence demonstration will explore profit margins, price positioning, and purchasing power.
Booper is participating for the second time in Spot, the retail industry’s back-to-school seminar, in Chantilly on October 1 and 2, 2026. We will be hosting a panel discussion titled “AI & Domain Expertise: Retail’s New Winning Combination,” featuring Ines Amor, who holds a Ph.D. in AI and data science, and Hugues Lafitte, who has over 10 years of experience in retail pricing.
Booper will be participating in NRF 2026: Retail's Big Show Europe, at Paris Expo Porte de Versailles, from September 15 to 17. Visit the team at booth I027 (Hall 6) to discuss AI-driven pricing and business rules, balancing margin, price image, and purchasing power.