Thin content on product and category pages: how to fix it

Thin content means pages with too little unique, useful text: a product page with one line copied from the supplier, a category page that is only a product grid, or hundreds of near-identical pages generated for every filter. Search engines struggle to rank these pages, and AI answer engines have nothing to quote from them.

Why it matters now

Google's September 2026 spam update finished rolling out on October 8, 2026, after almost two weeks. Google did not say which spam category it targeted, and only advised affected sites to review its spam policies. Those policies include scaled content abuse: many pages produced mainly to rank rather than to help people, whatever the method used to create them, AI included.

An online store is not at risk because it has many pages. It is at risk when many of those pages add nothing of their own. Thin pages also hurt on their own terms: they rarely rank, and they dilute the pages that should.

Where thin content hides in a store

• Supplier descriptions pasted unchanged, shared word for word with every other shop selling the same product.
• Category pages with a title and a product grid, and no introduction.
• Filter and sort URLs (colour, size, price, sorting) that are crawlable and indexable.
• Variant pages that differ only by a colour name or a size.
• Auto-generated text that swaps the product name into the same template on every page.

How to fix it

1. Prioritise. Start with the categories and products that bring traffic or sales, not the whole catalog at once.
2. Rewrite product descriptions with what only you can say: materials, sizing, how it fits or works, who it suits, what is in the box, care instructions.
3. Add a short intro to category pages: two or three sentences that help a shopper choose, placed above or below the grid.
4. Consolidate near-duplicates. Point variant and filter URLs to the main page with a canonical tag, or keep them out of the index.
5. Give every indexable page a unique title, as covered in product page titles and meta descriptions.
6. Support the text with facts machines can read, such as Product structured data.

Using AI to help write descriptions is not the problem in itself. Publishing hundreds of unreviewed, interchangeable texts is. Check every generated description against the real product before it goes live.

How Averifly helps

Averifly flags pages with fewer than 100 words of visible text, duplicate titles, product pages without a meaningful description, and canonical tags pointing elsewhere. See every check in our rules, then run a free scan to find the thin pages in your store.