GTIN is an internationally recognized unique product identifier, printed on the packaging next to the barcode. In a feed it lets systems place a product unambiguously instead of guessing from the name. It is sent in the gtin attribute, and according to Google it helps its systems categorize the product. Not every product has a GTIN, though — and you can’t invent one.
This is the last piece in the series on AI SEO for e-commerce, and the most technical of the bunch. It’s also the layer everything else sits on: without reliable product data, there’s no point worrying about copy or reviews.
Why the identifier matters more than it looks
Every retailer writes product names their own way. “Ultra Gloss Polish 16 oz,” “Ultra Gloss polishing compound, 1 pint,” and “ULTRA GLOSS 16” can all be the same product — and you won’t reliably tell from the name.
An identifier shrinks that problem by definition: it’s unique and internationally recognized, so it carries unambiguous information about the product no matter what anyone decided to call it. Google states that the GTIN is sent in the gtin attribute and helps its systems categorize and display products.
Three layers that need to agree
This is where online stores lose the most ground — and it isn’t a technology problem, it’s a maintenance problem.
| Aspekt | Layer | What usually goes wrong in it |
|---|---|---|
| Page | What the customer sees | Manual edits that never propagate anywhere else |
| Feed | The data file for Merchant Center and other shopping channels | Generated automatically, often from older fields |
| Structured data | The machine-readable description on the page | Left over from a previous version of the template |
When those three layers drift apart, you get a contradiction nobody can resolve. Keeping them in sync isn’t an explicit rule about feeds, but it follows from two documented obligations: the Shopping ads policies prohibit presenting products in a way that isn’t accurate and truthful, and for structured data Google requires a match with the visible content of the page.
Price is the most sensitive field of all — why that is, is covered in the article on pricing on your site.
How to get the data in order
-
Find out which products have a GTIN
It's on the packaging next to the barcode. Google states that not every product has one — own manufacturing, bundles, made-to-order.
-
Add the identifiers
GTIN into the gtin attribute. For products without one, the brand and the manufacturer's part number.
-
Make the data agree across layers
Price, availability, name, and specs have to say the same thing on the page, in the feed, and in structured data.
-
Vet third-party identifiers
For supplier GTINs, Google recommends verifying they are legitimate; if you can't find one, ask the manufacturer.
-
Check on a schedule
The feed runs automatically, pages change by hand. Walk a sample of products repeatedly.
What to do with products that have no GTIN
Not every assortment has one, and Google acknowledges that explicitly. It mostly applies to your own manufacturing, bundles made up of several items, and made-to-order production.
The answer is not to invent an identifier — that’s the worst possible option, because an incorrect GTIN can match your listing to an entirely different product. What works in practice instead is adding the brand and the manufacturer’s part number, so the product can be identified by at least that pair.
Four mistakes in product data
An invented or inherited GTIN
An incorrect identifier can categorize the product wrong. For identifiers from third-party sources, Google recommends verifying they are legitimate.
The feed and the page disagree
Two contradictory values for one thing. And presenting a product inaccurately is prohibited by the Shopping ads policies on top of that.
Structured data with nothing behind it
A value in structured data that isn’t visible on the page runs against the requirement to match the visible content.
Set the feed up and forget it
Prices, availability, and the assortment all change. Without a regular check, the layers drift apart on their own.
Where this guide stops
It doesn’t promise placement or recommendation. Consistent data is a precondition for reliable identification, not a guarantee your product turns up in a specific answer.
It doesn’t list the requirements of individual channels. Google Merchant Center, Bing Shopping, and Meta’s catalog each have their own formats and documentation; reciting them from memory would be unreliable.
It isn’t a Merchant Center walkthrough. This article is about the data, not about clicking through an interface.
What to take away
Product data is the least rewarding part of running a store: it’s invisible, nobody praises you for it, and it breaks quietly. Yet everything else sits on top of it — copy and reviews are worthless if the system doesn’t know which product it’s looking at.
The identifier helps with categorizing and recognizing the product; consistency across layers helps with how trustworthy the values are. Neither is difficult; both require maintenance.
Related reading: how to write the product page itself is covered in product pages for AI, what shopping assistants pull from product data in AI shopping agents, and how to handle reviews in reviews and ratings for AI.
Not sure whether your feed and your pages agree? This site is run by Sniper Design — an AI SEO audit walks your product data, your structured data, and whether the two line up, then names what to fix. Ready-made product and category page templates cut out the guesswork on the layout itself.
Sources: Google Merchant Center documentation on the GTIN identifier and the Shopping ads policies, current as of July 18, 2026.
