Competitive product matching: make price comparisons more reliable
March 15, 2026
Product matching is the foundation of competitive monitoring: it prevents companies from lowering prices based on a false price difference between two non-equivalent products. Cases of doubt are referred to business-unit review.
According to Diamart Group (2026), about 50% of French retailers still perform their price matching manually, while 94% collect competitive data.
Does your price monitoring really compare equivalent products? Just one incorrect match is enough to trigger an unnecessary price drop and blur your price image.
This article explains the causes of product-matching errors (variants, lots, price adjustments, marketplaces, out-of-stock items), how a competitor's price can be used as a reliable basis for decision-making, and how to assess the accuracy of a tool.
Why Product Matching Is the Critical Aspect of Competitive Monitoring
Product matching involves linking each price collected from a competitor to the equivalent product listing on your catalog. It is the foundation of data-driven pricing.
Without it, a price difference may simply be a difference between two different products. And a decision based on that difference is flawed, regardless of the quality of the data collected.
Percentage of French retailers surveyed
Source: Diamart Group, “Pricing Maturity of French Retailers,” 2026.
Manual matching isn't a sign of incompetence. With tens of thousands of listings and several competitors, it simply can't keep up with price changes.
The Cost of a False Match
A false match costs more than a missed match. This is the principle that should guide any system.
An unreconciled item triggers no action: information is lost. An incorrectly reconciled item triggers a decision—for example, a price reduction in response to a discrepancy that does not exist.
To give you an idea of the scale: a product that sells for €3, with 10,000 units sold per week, but is mistakenly priced at 30 cents for a smaller size, loses €3,000 in profit each week.
No one notices this, since thePrice Index is actually improving. That is why we are convinced that, for anything that feeds into an automated decision, accuracy takes precedence over coverage.
Exact match, equivalent, or substitute
An exact match refers to two identical products: same EAN, same brand, same volume.An equivalent product shares similar characteristics without being identical, such as a store brand compared to a national brand.
The substitute meets the same customer need. Each level triggers a different decision rule: pricing can be aligned with an identical product, but rarely with a substitute.
Reliable matching is essential for a fair analysis of competitors' prices.
The Top Causes of Matching Errors (and How to Spot Them)
These pitfalls are best addressed at the source—when the price is collected—rather than at the end of the process in a spreadsheet.
Variants, Lots, and Units
A single difference in size, color, capacity, or model year is enough to change the price: a 1 TB drive is not the same as a 2 TB drive. Bundles (x2, “+1 free”) distort the list price.
The trick is the same everywhere: convert each offer to a unit price (per kilogram, per liter, per item) before comparing. An incorrect conversion—such as reading grams as kilograms—results in a spectacularly false difference.
Redflation: The Hidden Pitfall of Matching
As of July1, 2024, large grocery stores with more than 400 m² must display, for two months, products whose quantity is decreasing while the price remains the same or increases (INC).
When it comes to matching, there are two pitfalls. According to GS1 rules, a change in volume requires a new EAN: matching by code results in the product being lost.
A comparison based on product descriptions, on the other hand, risks pitting the old format against the new one. Only a comparison based on unit price holds up, and that is often where a competitor’s actual price increase is hidden.
Incomplete references
Without an EAN or manufacturer part number, matching is based on the brand, model, attributes, and image. The weaker the signals, the more essential human verification becomes.
Among our customers, the EAN is often missing or varies from one retailer to another. The matching process therefore also relies on text recognition (NLP) and image recognition, which link two products even when their names do not match.
In a multi-brand franchise group (discount, organic, frozen foods) that transitioned from Excel to a tool-based management system, the product mapping process allows the company to start with a clean slate for each six-week pricing cycle.
On the data collection side, Booper gathers prices online (websites, marketplaces, apps) itself. Store surveys and panelist data provided by the retailer are integrated into the same cross-reference table.
Team Booper: Competitive Intelligence and Product Matching Projects in the DIY and Multi-Brand Retail Sectors
There remains one pitfall that the recognition of the wording alone does not resolve: the nature of the offer.
Marketplaces, Promotions, and Availability
A third-party seller on a marketplace does not offer the same shipping rates, the same warranty, or always actual inventory. Its price cannot be directly compared to that of the official seller.
A promotion that ends tomorrow or a promo code applied to the shopping cart says nothing about the regular price. And an aggressive price on a product that’s out of stock is a decoy: matching that price is tantamount to selling at a loss against a competitor who isn’t making any sales.
Average out-of-stock rates on store shelves for fast-moving consumer goods in early 2023: enough to skew a monitoring system that does not filter out stockouts.
(LSA-NielsenIQ Barometer, 2023)
A price listed for an unavailable product should therefore not be deleted, but marked as such: it still provides insight into the competitor's strategy.
How Automated Product Matching Works
Making product matching more reliable doesn't mean trying to match everything at once. Flawless matching for the most frequently compared products and direct competitors is better than rough, comprehensive coverage.
The quality of your own product listing (EAN, brand, model, capacity) determines the quality of the external match. The internal database must be cleaned up before looking elsewhere.
The matching process then cross-references various signals (codes, labels, images, attributes) and assigns a confidence level to each match. That is what our product-matching software does.
Our belief: Matching isn't a one-time project; it's an asset that requires ongoing maintenance. Every correction made by a category manager must feed into the system; otherwise, the same error will recur in the next cycle.
Collection
Online prices collected by the tool, in-store price data, and panelist data provided by the retailer.
Matching
Each price is linked to the correct product listing by name, image, and attributes.
Index
The spreads are widening in the " Price Index" segment, weighted by sales volume.
Simulation
The impact of an adjustment on sales, margin, and inventory is simulated before taking action.
Business Validation
The pricing or category team either approves the recommendation or rejects it.
A false match that passes Step 2 affects all subsequent steps: the index, the simulation, and the recommendation are then based on a discrepancy that does not exist.
Manual matching or AI-powered matching
The manual provides detailed information on a few hundred products but does not cover the full range. The AI covers the full range, but it confidently makes mistakes if no one reviews its questionable cases.
The right balance, therefore, is not one or the other: AI handles the bulk of the work, humans resolve ambiguous cases, and their corrections are used to retrain the system.
Case of multi-brand and multi-collection groups (fashion, specialty)
A single product may be available in multiple colors, materials, or seasonal editions, each with a different price point.
The repository should therefore list the brand and the collection as top-level attributes, and the scope of the monitoring should be broken down by brand rather than by generic category.
Large-scale product matching and chaining, in 1 minute
Read the video transcript
Thousands of products in your store, thousands at your competitors'. Which one matches which?
That's what product matching is all about! When done manually, you compare, verify, and decide.
Line by line. In this example: four products processed… out of 12,400.
Phew! First clue: the barcode.
When the same code appears on both sides, the match is immediate. Easy!
But your private-label brands don't have a common code. So the AI looks at the product: its outline, its proportions, and its size.
Third clue: the wording. Volume, color, scent: the AI reads the variation.
Each match is assigned a confidence score. What about borderline cases?
Set aside, to be checked. With Booper, you can scale up: tens of thousands of products processed, whereas manually, we're still only on the first few lines.
The result: faster processing, fewer errors, and targeted human review. Your teams now only check what needs to be checked.
Key takeaway: AI handles the bulk of the work; your teams make the final call on uncertain cases. Product matching, on a large scale!
Exact or Equivalent Match: What Are the Pricing Rules?
When prices match exactly, price matching is often the norm, because customers compare the same product across different websites. There is still a choice to be made: whether or not to match the price of a competitor that is out of stock or a questionable third-party seller.
When comparing equivalent products, we don't look for price parity. A price difference is justified by the brand, service, or warranty, and it is managed rather than corrected.
| Product Type | Price sensitivity | Matching Rule | Recommended Action |
|---|---|---|---|
| KVI (Best Sellers) | High | Correct | Strict alignment |
| Niche products (long tail) | Bass | Large | Margin Preserved |
| Seasonal Products | High | Exact or Approximate | Dynamic Alignment |
The effort required for matching is proportional to what is at stake: absolute precision regarding the products to monitor as a priority, and greater flexibility regarding the long tail.
Measures to Combat Match-Fixing: What Needs to Be in Place Before Adjusting Prices
Before approving a new price, four factors must align: the product (brand, model, capacity), the offer (seller, shipping, warranty), the promotional terms, and the product's condition.
One extra letter in a product code, a charge for shipping, or a refurbished item is enough to make two prices incomparable.
These checks are not performed manually, line by line. The tool must automatically filter out questionable bids and submit only ambiguous cases to the team.
How to Evaluate the Accuracy of a Product Matching Tool
A high matching rate means nothing without its error rate. A 95% coverage rate with false matches on image products is worth less than a 70% coverage rate with no errors.
Accuracy is measured based on a random sample of automatically validated matches, not on the examples chosen for the demonstration. It is rechecked after each software update.
Next comes the coverage of strategic products, which must be fully reconciled, followed by the business impact. If the margin is falling even though the index looks good, first check the reconciliation.
Governance: Who Approves What?
The principle is based on three levels of confidence. Confirmed matches drive automated decisions; ambiguous cases are reviewed; and cases involving strong doubts are blocked.
The pricing manager sets these thresholds, the category manager resolves ambiguous cases, and the data team ensures data flows. This is the organizational structure adopted by a national home improvement retailer: competitors’ prices feed into the index, and the central team approves each price adjustment.
Competitive prices that help you make decisions, not just sit on a shelf.
Competitors' prices are imported directly into the tool and matched to the retailer's products by cloning and linking SKUs. The price indices are up to date, and each price alignment is simulated before being approved by the team.
- Issue
- Track national competitors across a very extensive product catalog, store by store.
- What Has Changed
- Price indices, cloning and chaining matching, and centrally validated recommendations.
platform availability in 2025, with the platform having been in production since 2022.
Pricing Department · National home improvement chain. Case study written by Booper based on an actual client project; the client has been anonymized.
The frequency of the review is just as important as the quality of the initial comparison: the most frequently compared products are monitored daily, while the rest are monitored less frequently.
Improving the Reliability of Product Matching: Where to Start
Start small, but aim for accuracy on the products that define your pricing strategy. You earn your teams' trust through verifiable results, not through full automation in the first month.
Data quality is an ongoing challenge, as competitors change their product descriptions, formats, and batches. That’s what turns competitor price monitoring into a margin driver.
See also: Building a Comprehensive Price Monitoring System.
See also in this feature
- Web Scraping of Retail Prices: Definition and Legal Framework
- BOOPER vs. Prisync: Which Tool Should You Use to Manage Your Retail Prices?
- Monitoring Competitors' Prices: How to Choose the Right Tool
- Excel's Limitations for Monitoring Competitors' Prices: The Threshold
FAQ
Product matching, or product mapping, links each your catalog SKU to identical or equivalent products sold by your competitors.
It is used to compare prices on a reliable basis. Without it, price monitoring produces false discrepancies, posing a direct risk to profit margins and price reputation.
It processes various types of data: EAN codes and manufacturer part numbers, labels analyzed using natural language processing, images, and attributes (brand, volume, format).
Each match is assigned a confidence level. The most reliable matches inform decisions, while ambiguous cases are reviewed by a human.
The manual is reliable on a small scale but cannot keep up with changes across thousands of product listings. AI scales up and links products with different descriptions.
The two complement each other: AI processes the bulk of the data, while humans resolve the borderline cases, and their corrections improve the next round of matching.
Randomly select a sample of automatically validated matches and check them one by one. The error rate measured this way is more meaningful than any advertised coverage rate.
Focus first on the errors in your most-compared products: those are the ones that trigger unwanted price drops.
An exact match refers to two identical products (same EAN, same model). An equivalent product has similar characteristics, while a substitute simply fulfills the same need.
The decision rule changes depending on the level of proximity: alignment is truly justified only when there is an exact match.
MonitorPrice Index 's performance relative to competitors, respond to price discrepancies on the most frequently compared products, compare product lines, and identify items missing from your own inventory or from competitors' inventories.
In any case, the value comes from the decision made afterward, not from the number of awards received.
By converting each offer to a unit price (per kilogram, per liter, per item) before making any comparisons, and by identifying phrases such as “x2,” “+ 1 free,” or crossed-out prices.
A change in package size results in a new EAN: without a unit price comparison, the old and new sizes get mixed up.
Determine what the tool can do on its own. A definite match can trigger automatic repricing; an uncertain match is reviewed; a questionable match is blocked.
This rule prevents an algorithm from lowering prices based on a tenuous correlation.
A comprehensive competitor pricing monitoring system is built on five inseparable components: data collection, product matching, alerts, reporting, and governance. If even one of these components is missing, the system becomes ineffective.
Data collection is no longer the weak link: according to Diamart Group (2026), 94% of French retailers collect competitive data, but about half still perform the matching manually.
Coverage, timeliness, and compliance are prerequisites for a competitor price monitoring tool. Its value depends on the accuracy of product matching and on linking each price to a traceable decision, yet approximately half of French retailers still perform this matching manually (Diamart Group, 2026).
The listed cost is only part of the total cost. Manually reconciling discrepancies, maintaining in-house development, and making decisions based on unreliable data often end up costing more than the subscription.
A successful pilot test proves almost nothing: on a small scale, human attention masks the flaws in matching, alerts, and frequency that become immediately apparent on a large scale.
Systems that are in place advance in stages, but no stage is mandatory: we move on to the next one when maintaining the previous one costs more than moving beyond it. And what scales up is not the number of employees, but the traceability of decisions.