Product Matching

Product matching is the process of identifying the same product across different stores, catalogs or marketplaces so listings can be compared. It relies on identifiers such as GTIN and MPN, or on titles, attributes and variant details when those are missing. Price comparisons are only as reliable as the matches.

Why matching comes before any price comparison

A competitor price is useful only when it belongs to the product you actually sell. If a store compares its 1 kg bag of coffee with a rival’s 500 g bag, the rival looks cheaper per listing but is far more expensive per gram. Any repricing rule built on that pair would cut the price for no reason.

Matching turns a list of competitor pages into a clean set of pairs: your product, their product, and the confidence that they are the same item. Price tracking, price indexes and market rankings all sit on top of that set.

The identifiers that make matches reliable

Structured identifiers are the strongest evidence that two listings describe one product. The most common are:

  • GTIN (Global Trade Item Number): the barcode family that includes UPC, EAN and ISBN. A GTIN is assigned to a specific product version, so a match on it is usually strong.
  • MPN (Manufacturer Part Number): the code the manufacturer uses for the item. Combined with the brand name, it is often unique.
  • SKU (Stock Keeping Unit): a code a store creates for its own inventory. SKUs differ between stores, so they help inside your catalog but rarely link you to a competitor.
  • Brand plus model name: weaker than a code, but useful for products that have no barcode, such as handmade or private-label goods.

Identifiers are not perfect. Some sellers leave them blank, reuse one barcode for several variants, or enter a typo. Treat an identifier match as strong evidence, not proof, and check the attributes when something looks odd.

Matching without codes: titles and attributes

When codes are missing, matching falls back on the text and data on the page. Titles are normalized by removing filler words, standardizing units and splitting out the brand. Then key attributes are compared: size, capacity, color, material, pack count and model year.

Attribute checks catch the most expensive errors. Two vacuum cleaner listings may share a long title, but one includes a spare battery and the other does not. A careful match compares what is in the box, not just the name on the box. Images can support the decision, but a photo alone rarely separates close versions.

Variants, bundles and pack sizes

Products with variants need matching at the variant level. A T-shirt sold in five sizes and four colors is twenty items, and stores may price some of them differently. Matching only the parent product hides those differences. Stores can use variant detection to spot which products in their catalog have variants before pairing them.

Bundles and multipacks need a different treatment. If a rival sells a 6-pack and you sell single units, the pair is not an exact match, but it can still be compared on unit price. Mark these pairs as equivalent rather than identical, so reports do not mix them with true one-to-one matches.

A worked matching example

For example, assume an outdoor store sells a 2-person tent at 189 dollars and finds three competitor listings with similar titles. Listing A has the same MPN and the same color, priced at 179 dollars. Listing B has the same model name but describes a 3-person version at 219 dollars. Listing C has no codes and a vague title, priced at 149 dollars.

  1. Accept listing A as an exact match, because both identifier and attributes agree.
  2. Reject listing B, since the capacity differs and the price reflects a larger product.
  3. Hold listing C for review. A 40-dollar gap on a vague listing is more often a different or used product than a real bargain.

With only listing A counted, the store is 10 dollars above the market for this tent. Counting C by mistake would have suggested a 40-dollar gap and an unnecessary price cut. These numbers are assumed for illustration.

Common matching errors

  • Matching a refurbished, used or open-box listing to a new product
  • Ignoring pack size, so a single unit is compared with a multipack
  • Pairing an older model year with the current one
  • Trusting a shared title when capacity or included accessories differ
  • Leaving old matches in place after a competitor changes the product on the page

A useful warning sign is an extreme price gap. When one competitor appears far cheaper or far more expensive than everyone else, check the match before reacting to the price.

Keeping matches accurate over time

Matches are not a one-time job. Competitors rename pages, swap products behind the same address and remove listings. Review a sample of matches regularly, starting with the products that carry the most revenue. Record why each match was accepted, such as identifier, attributes or manual review, so problems are easier to trace.

Finding candidate pages is a separate step from confirming them. BlackAha’s competitor discovery finds competitor product URLs automatically, and the store can then review the candidates that matter most before using them in pricing decisions.

Key takeaways

  • Product matching links the same item across stores so prices can be compared fairly.
  • GTIN and brand plus MPN are strong signals; SKUs are store-specific.
  • Match at the variant level and treat multipacks as equivalents, not exact matches.
  • An extreme price gap is often a matching error, not a market signal.
  • Review matches regularly, because competitor pages change over time.

Frequently asked questions

What is the most reliable identifier for product matching?

A GTIN, such as a UPC or EAN barcode, is usually the strongest single signal because it is tied to a specific product version. A brand plus MPN is a good second choice. Neither is perfect, so attributes should still agree.

Can products be matched without barcodes?

Yes. Titles, brand, model name and attributes such as size, color and pack count can identify the same product. These matches carry more risk, so high-value pairs deserve a manual check.

Should variants be matched separately?

Yes, when variants can have different prices. Matching only the parent product hides price differences between sizes or colors and can make a store look cheaper or dearer than it really is.

How often should product matches be reviewed?

Review high-revenue matches more often, and check any pair that shows a sudden, extreme price gap. Competitor pages change, so a match that was correct last quarter may not be correct today.

Related terms

In BlackAha, Price Tracking helps you monitor competitor prices and stock status automatically.

Back to the Pricing Dictionary