How to fix an invalid product category in Merchant Center

What “invalid product category” actually means

Google’s own page for this issue is short and worth reading exactly. It says that if the attribute is not provided, Google will use internal signals to determine the correct category; that if you do provide it, you should use the predefined values; and that invalid values will be rejected and may prevent your products from appearing.

So there are two different situations, and only one is a problem. A product with no google_product_category is categorised automatically. A product with a value Google does not recognise has sent something wrong, and that value is rejected. Adding a category to every product is not required; fixing the bad ones is.

Deleting a bad category is a valid fix. Sending a wrong one is not.

Where invalid values come from

A value is valid when it is one entry of the Google product taxonomy, given as its ID or its full path. In a real feed, it usually stops being one in one of these ways:

1. A typo or an invented path. A path has to match the taxonomy exactly, level by level. Apparel > Shoes reads fine to a person and does not exist: the real branch is Apparel & Accessories > Shoes.
2. Your own product_type in the wrong field. Store navigation such as Women > Summer > Dresses belongs in product_type, which you define yourself. In google_product_category it is simply not a Google category.
3. Another platform’s taxonomy. Shopify’s product category and Microsoft’s own taxonomy are separate trees. A value from either, copied into the Google column, will not match — see Shopify.
4. Two values in one field. Google asks for one category, and for either the ID or the path — but not both. 2271 - Apparel & Accessories > Clothing > Dresses is a copy of the taxonomy file’s line, not a value.

Finding every bad row, not just the first one

Merchant Center lists the affected products under the issue. For a small feed, fix them there. For a large one, export the feed and check the column in bulk, because the cause is usually systematic — one import rule or one app writing the wrong thing into every product.

A quick way to triage a spreadsheet export:

• Values that contain a hyphen between a number and a path are the “both” case.
• Values without any > that are not purely digits are rarely valid, apart from top-level categories such as Apparel & Accessories.
• Values that repeat your store’s own menu labels are product types in the wrong column.

Replacing bad values with real IDs

Google’s specification recommends the numeric ID over the path, and for a repair job that is the sensible target: an ID is short, has no spelling to get wrong, and survives copy-paste. Two routes:

• A handful of products: look each one up in the Category Finder, which returns the ID, the full path and a confidence score from the product title.
• The whole feed: run the export through the Bulk Mapper. With an id column mapped, it writes supplemental-feed.csv — just id and google_product_category — for the rows it is confident about, and a needs-review.csv for the rest.

Upload the supplemental file as a supplemental data source linked to your primary one; values are matched on product id. The guide to overriding the category with a supplemental feed walks through that upload.

When removing the value is the right fix

If a value came from a broken import and you have no better one, deleting it is legitimate: the product falls back to Google’s automatic category. What you should not do is swap an invalid value for a valid but wrong one to make the warning go away. A wrong category is quiet — nothing flags it — and it is worse than none for the few products where the category decides which rules apply; see products that need a specific category.

Keeping it fixed

Fix the source, not just the feed: the app, import template or feed rule that wrote the bad value will write it again on the next sync. Keep categories as IDs, keep store navigation in product_type, and re-check the column when your import process changes.