Duplicate Product Descriptions: Does Google Penalise Them?

You have 2,000 products and every one of them carries the exact text that arrived in the supplier feed. Seven other shops imported the same feed and have the same text. The question is always the same: will Google penalise me for it?

No. The outcome is still bad, it just has a different name.

Penalty versus filtering — the distinction that matters

A manual action is a specific decision you can see in Google Search Console under Manual actions. Copied feed descriptions almost never trigger one. Nobody is removing your shop from the index.

What actually happens: Google picks one version of the repeated text as the main one and either pushes the rest down or leaves them out of the index entirely. In Search Console it shows up like this:

• Duplicate, Google chose different canonical than user — your URL has been folded into another one in the index.
• Crawled – currently not indexed — the crawler saw the page and found no reason to include it.
• Discovered – currently not indexed — the crawler never even fetched it, because the site has thousands of near-identical URLs and the crawl budget gets spent elsewhere.

The third one hurts most on large shops. You do not have a ranking problem. You have a problem where half your catalogue does not exist in search at all.

Where duplicate descriptions come from

In practice I see four sources, and each one needs a different fix.

The supplier feed. The text is identical across every retailer selling that item. Your page has no reason to beat a shop with a stronger domain selling the same product with the same words.

Variants of one product. A t-shirt in six colours and four sizes is 24 URLs sharing one description. That is not a content problem, it is a structure problem.

Copy-paste between your own products. Thirty models of garden hose, each with the same paragraph about quality materials, differing only in the spec table.

Empty descriptions. Zero characters is not duplication, but for search it is worse. The page is not about anything.

How to find out how many products are affected

Three steps you can do today.

1. Search Console, Page indexing report. Look at the URL counts in the states above. If you have 3,000 products and 900 indexed, you know where the problem sits. Click a specific URL and the URL Inspection tool will show you which page Google picked as canonical.

2. Search a snippet in quotes. Take the first sentence of a typical product description and paste it into Google inside quotation marks. If ten shops come back with the same sentence, you have your answer. Run the same test restricted to your own domain — that tells you how many of your own pages carry identical text.

3. Export the catalogue. Pull descriptions into a spreadsheet and sort by character count. Products under roughly 200 characters, and groups of products with exactly the same text length, are your candidates for review.

Seonal does this step for you: I crawl the public pages of the shop, compare the texts against each other and return a list of URLs with identical or near-identical descriptions, including how many times each block of text repeats across the site.

You do not have to rewrite everything

With 3,000 products, rewriting every description is a project measured in months. Starting alphabetically from the top makes no sense.

An order that does:

1. Products that already get impressions in Search Console but sit at position 8 to 20, or have weak CTR. Google knows them, it just does not trust them yet. This is the fastest movement you will get.
2. Products with the highest margin or revenue. Best return per hour of work.
3. Products that are not indexed at all. More work here, because first you have to give the crawler a reason to fetch them.
4. Everything else. Possibly never, if it is stock people only ever search by manufacturer part number.

If you are not sure which products belong in group one, Search Console data answers that more precisely than a guess.

What a non-copied product description contains

You do not need prose. You need information the supplier text does not have.

• What it is for and who it is for. One sentence describing a situation, not a feature list. "For a balcony up to two metres wide" beats "a universal solution".
• Specs in the body text, not only in the table. Dimensions, material, output, compatibility. People search for "monitor arm for 27 inch", not "monitor arm".
• How it differs from the similar model you also stock. The supplier will never write this, because they do not know your range. It also solves duplication between your own products.
• Two or three questions that reach your support line. Does it fit the older model? Do I need an adapter? How long is delivery?
• What is in the box and what is not. The most common reason for returns, and unique text at the same time.

Three hundred to five hundred characters of that kind of text on top of the inherited description is enough to stop the page being a copy. You do not have to delete the supplier text — it just should not be the only thing on the page.

A product description and a meta description in search results are two different things. The first is page content, the second is an invitation to click. Both tend to be duplicated and both need fixing, but separately.

Fix variants with canonicals, not with rewriting

If you sell one t-shirt in six colours and each colour has its own URL, writing six different descriptions of the same t-shirt is wasted effort. This is where a canonical tag belongs, pointing the variants at one main URL.

The exception: when people genuinely search for the colour as its own query ("white Adidas trainers"), that variant deserves its own indexable page with its own text. You can check the decision in Search Console — if the variant collects impressions on its own queries, leave it alone.

Parameter URLs with filters (?colour=black&size=42) are a separate chapter. As a rule they do not belong in the index at all.

Duplicate descriptions rarely travel alone

A shop with identical product descriptions almost always has duplicate title tags and missing image alt text too. It all comes out of the same import.

So when you fix, do not fix only the body copy. Check at the same time:

• the page title — does it carry brand, model and the spec people actually search by?
• product structured data — price, availability, reviews;
• internal links to related products and categories, so the crawler can reach deep pages in the first place.

How to check whether the rewriting did anything

Without measurement it is faith. The method that works:

Write down the deployment date. Four weeks later, compare 28 days before against 28 days after in Search Console — but only for the URLs the fix touched, not for the whole site, because at site level the numbers get moved by seasonality and everything else.

Watch three things: the number of indexed products, the number of queries the page appears for, and position. On product pages impressions usually move first — the page starts showing up for queries it was absent from entirely.

And allow time. Positions move in weeks, not days. Anyone telling you otherwise is lying. More on that in the article on measuring SEO impact.

What I do with this

I crawl the public pages of the shop, find products with duplicate or very short descriptions, and for the priority findings I write the finished text — not a note saying "add unique content". You read it, approve it or edit it, and it goes live through the WordPress plugin, the API, or a CSV export.

Then I compare 28 days before and after deployment so you can see what specifically moved.

Rewriting 3,000 descriptions at once makes no sense even when someone can produce them fast. What makes sense is rewriting the 150 that already get impressions and are only missing a reason for Google to prefer them over seven shops running the same text.