In short
- The copied supplier description is a problem, not the main one. You fix it, positions do not move, and teams get discouraged for the wrong reasons.
- Your product page is up against your own category page on the same search intent, and the category almost always wins because it receives all your internal links.
- It is also up against your marketplace listing, selling the same product from a far stronger domain.
- Variants multiply the problem. One dress in five colours and six sizes can produce thirty near-identical pages that cancel each other out.
The usual diagnosis fits in a sentence: your product pages reuse the supplier’s text, so they are duplicates, so they do not rank. That is true and insufficient. We regularly see fully rewritten catalogues, at considerable expense, whose positions have not moved an inch.
When rewriting produces nothing, the problem was not there.
The three competitors you had not counted
A product page rarely fights other brands’ pages. It first fights three pages that all either belong to you or represent you.
| Competitor | Why it wins | What to do about it |
|---|---|---|
| Your category page | It receives nearly all your internal links, from the menu and from every product page | Decide which of the two should rank, and link accordingly |
| Your marketplace listing | The domain is vastly stronger, and the listing often carries the same title | Differentiate the title and the angle, do not duplicate word for word |
| Your own variants | Five colours and six sizes can produce thirty near-identical pages | One product page, variants not separately indexable |
| The manufacturer’s site | Same text, higher authority, earlier publication | The only case where rewriting the description genuinely solves something |
The mistake that costs the most Launching a full catalogue rewrite before checking which of your pages already ranks on the target query. If it is your category page, rewriting the product page will change nothing: you will have paid for copy in a fight that happens elsewhere.
The check that should precede any rewrite
It takes a few minutes per query and it changes the priorities.
- Take ten products that matter and, for each, the query a customer would actually type. Not the internal reference name: what a person writes.
- Search that query with your domain name added. You will see which of your pages is retained. Nine times out of ten it is the category, or a marketplace listing.
- Check the Search Console performance report for that query. If it shows impressions on the category and zero on the product page, the subject is internal linking, not copywriting.
- Count your indexed variant pages. A site search on a product name tells you in ten seconds whether you have one page or thirty.
What this check produces, almost every time: a much shorter list of pages to rewrite, and an internal linking project nobody saw coming.
And when the arbitrage points to the category, the next project is not a copywriting one at all: it is the order that category displays its products in, which decides what the visitor sees first.
What makes a product page genuinely useful
When the product page is indeed the one to rank, the copywriting work makes sense. It does not consist of making the text longer.
What distinguishes the product pages that rank, in our experience, comes down to three things. They answer the questions asked before purchase rather than the product’s characteristics: does it run large, does the fabric itch, how long does assembly take. They carry concrete, verifiable data, dimensions, weight, composition, care, which is exactly what engines and assistants know how to extract. And they say who the product is not for, which raises conversion and cuts returns.
That last point is counter-intuitive and it is the one that pays most. A page that turns away the wrong buyers sells less and earns more, and it limits the cost we cover in our article on the real cost of returns.
The options, and what they are worth
| Option | What you get | Who it suits |
|---|---|---|
| Arbitrate category against product page, then link | The highest-return move, and it costs only decisions. Often gains within weeks | Everyone, before writing a line |
| Sort out the variants | Removes massive internal competition on fashion and equipment catalogues | Catalogues with variants, so most of them |
| Rewrite the pages that matter | Effective once the two previous points are handled, and on a scope limited to products with real stakes | Thirty to fifty references, not three thousand |
| Generate the whole catalogue automatically | Produces decent, generic text. Without prior arbitrage you get three thousand better-written pages still competing | As support, never as strategy |
Tooling
Two needs: spotting genuinely thin pages across a large catalogue, which is measurement work, then deciding what to add to them.
On PrestaShop we deploy the modules published by Datafirefly Limited, our agency’s sister company: Thin Content to identify pages that are too thin, rank them and suggest the enrichments to add, and Smart Content to personalise the display per visitor segment, selecting the most relevant content block in real time. Neither module writes your product descriptions for you. One-off purchase · 12 months of updates.
Method beats tooling, and the order is decisive here: generating copy before arbitrating between category and product page amounts to writing better pages that will keep cancelling each other out.
What this says more broadly
Catalogue search is treated as a volume problem: more pages, more text, more keywords. The opposite has worked for a long time.
A catalogue that ranks is one where each search intent has one single page designated to serve it, and where all the others point at it. That is architecture work, not copywriting, and it follows the same logic as the content audit we run within our acquisition practice: decide what should exist, before deciding what to write.
Sources
The verifiable claims in this article link to their primary source, consulted on 10 August 2026. We do not cite a source we have not read.
- Google Search Central, How to Specify a Canonical with rel=”canonical” and Other Methods. View
- Google Search Central, Ecommerce URL Structure Best Practices. View
- Google Search Central, Spam policies for Google web search. View
- Google Search Console Help, Performance report. View
- Datafirefly Limited, Thin Content Detector, product page. View
FAQ
Do I need to rewrite every supplier description?
No, and that is the classic budget trap. Check first which page of your site ranks on the target query. If it is your category, rewriting the product page changes nothing. Rewriting is justified on products with real stakes, once the arbitrage is done.
Should my variants have their own page?
Rarely. Unless customers genuinely search for the variant, for instance a colour that has become a product name, one page with variants that are not separately indexable beats thirty near-identical pages.
Does being on a marketplace penalise me?
It does not penalise you, it competes with you. The listing benefits from a stronger domain and often carries the same title. The answer is not to leave the marketplace, it is to stop duplicating the same title and the same angle.
Is AI-generated copy penalised?
The tool is not the problem, the result is: generic text answers no question and does not rank, whether a person or a machine wrote it. The useful test stays the same: does it bring information the customer cannot get elsewhere.
Can Dotsland help?
Yes. Arbitrating between category pages and product pages, putting internal linking back in order, handling variants, then rewriting the references that matter. It is part of our acquisition practice. Let’s talk, or start with the check on ten products.