Duplicate content in ecommerce: variants, filters, canonicals

An online store produces duplicate content almost by itself. Every variant, every filter, every sort order can end up as its own URL. This is rarely a penalty and usually a waste: Google crawls dozens of near-identical pages and splits signals like links, clicks and internal references across several URLs instead of concentrating them on one. This article covers where the duplicates come from, which fix belongs to which case, and how to verify the state of things in Search Console.

What duplicate content actually costs you

Google detects near-identical pages and picks one of them as the canonical URL. Your canonical declaration is a strong signal, not a command. When internal links, sitemap and canonical contradict each other, Google decides, and sometimes it decides in favour of the filter URL instead of your category page. Google says as much in its documentation on consolidating duplicate URLs: if you do not specify a preference, Google will identify the version it considers best.

The consequences are undramatic but they cost visibility:

• Crawling spreads sideways. The crawler collects tens of thousands of parameter URLs and reaches new products and updated category copy less often.
• Signals fragment. An external link to ?color=blue does not strengthen your main page when the relationship is unclear.
• The wrong URL ends up in the index. Visitors land on a pre-filtered list instead of the category page you actually optimized.
• Your own pages compete. Two nearly identical categories for the same term do not double your chances, they produce two mediocre positions.

A manual action only becomes a risk when copied content is published at scale to manipulate search results. Variant and filter URLs are not that.

Where the duplicates come from

Variants

Colour, size, length, quantity. Many platforms append the selection as a parameter on the product URL. The text stays the same; sometimes the image and price change.

Filters and facets

?brand=, ?size=, ?price_max= and every combination of them. Five filters with five values each add up to thousands of URLs, almost all showing the same products in a different order.

Sorting and view options

?sort=price-asc, ?view=list, ?per_page=48. These parameters never change the content, only the presentation. This is the purest form of duplication.

Pagination

Pages 2, 3 and 4 of a category. Not a real duplicate, but frequently mishandled: pointing every subsequent page at page one with a canonical hides the products on the later pages.

Internal search results

/search?q=... generates unlimited pages that are empty or randomly filled depending on the input. They do not belong in the index.

Tracking parameters

?utm_source=newsletter, click IDs from ad platforms, affiliate parameters. Technically the same page, a new URL for the crawler.

One product in several categories

Some systems expose the same product under several paths. Shopify handles this internally with a canonical to the /products/ URL; in other systems you have to verify it.

Host and spelling variants

example.com and www.example.com, http and https, with and without a trailing slash, mixed case in the path, /index.php alongside /. Every variant that serves a page instead of redirecting is a separate URL.

Print, preview and staging views

Print versions, preview links, staging domains without protection. Classics, because nobody published them on purpose.

Textual duplicates

The same supplier text on hundreds of stores, identical descriptions across every variant, category copy where only the category word was swapped. Technically correct, editorially interchangeable. If you draft with generative tools, the rules that keep this from getting worse are in the article on AI product descriptions.

The tools and what they really do

Two pieces of advice that get passed around long after they expired. First: the URL Parameters tool in Search Console no longer exists. Google announced its deprecation in March 2022, because only a tiny fraction of the configurations stored in it were still useful for crawling and Google handles parameters automatically now. Today you steer parameters through canonicals, internal links and robots.txt. Second: rel="next" and rel="prev" are no longer used by Google. The documentation says plainly that Google no longer uses these tags, although the links may still be used by other search engines. The same page also tells you to give each paginated page its own canonical URL and not to use page one as the canonical for the whole sequence — see Google's guidance on pagination and incremental page loading.

And one rule worth writing on the wall: never combine noindex with a canonical pointing at a different page. Google's own wording is that it does not recommend using noindex to prevent selection of a canonical page within a single site, because it blocks the page from Search completely, and that rel="canonical" is the preferred solution. Pick one mechanism per URL.

Decision table: which case needs which fix

Rule of thumb: if a human being would search for the page in that exact state, it belongs in the index. If not, it does not, and the only remaining question is whether noindex or a crawl block is the right tool.

Step by step: clean up in seven steps

1. Take stock. Open the page indexing report in Search Console and read the not-indexed statuses. Their exact labels in the English interface are "Duplicate without user-selected canonical", "Duplicate, Google chose different canonical than user" and "Alternate page with proper canonical tag". The first is a task, the third is the normal state. If you also see a large block of "Crawled - currently not indexed", work through that status in parallel, because the causes overlap.
2. Collect the parameters. Write down every parameter your store generates: filters, sorting, variants, tracking, pagination. Your server logs or a crawler's URL list show what Google actually fetches.
3. Classify each parameter. Does it change the content, only the presentation, or nothing at all? That determines which row of the table above applies.
4. Check demand before you delete. For filters and variants, verify whether there are searches for them. In Search Console, the Queries view tells you what the page already earns impressions for today.
5. Unify canonicals. Every indexable page names itself; every duplicate variant names the main page. Use absolute URLs and the same spelling as in the sitemap and internal links.
6. Straighten internal links. Sitemap, navigation, breadcrumbs and product grids may only point at canonical URLs. When canonical and internal links contradict each other, the link often wins. The wider structure sits in the ecommerce SEO guide.
7. Measure again. After two to four weeks, revisit the same reports: has the number of duplicate statuses dropped, are the right URLs indexed, do category pages rank instead of filter URLs? For individual URLs, URL Inspection shows the canonical Google selected.

Five mistakes that come up most often

• robots.txt block plus noindex. Google cannot read the noindex. Choose one.
• Every paginated page canonical to page one. You are declaring the products from page two onwards unimportant.
• Canonical to a redirected or unreachable URL. The signal gets ignored.
• Canonical to the homepage. This gets set on product pages more often than you would think, and the result is that no product page ranks at all.
• Redirect chains. http://example.com to https://example.com to https://www.example.com to the target. Shorten every chain; one hop is enough.

What platforms already handle for you

Shopify sets canonicals for products reachable through collection paths and maintains its own robots.txt, editable through the robots.txt.liquid template. The specifics of variants, where Shopify's model differs from most systems, are covered in Shopify product variants and SEO, and the platform settings in the Shopify SEO guide. Lightspeed eCom also sets canonicals for products assigned to several categories, but the available settings differ by version, so check what your backend offers instead of following a generic click path.

In both cases the same caveat applies: automatic canonicals cover product URLs, not the filter and sort parameters your navigation or a filter app produces. That is where the volume comes from. So after every new app and every theme update, check whether new parameter URLs suddenly appear in the index.

If you want to keep an eye on this continuously, the SEO audit in ecomando lists which URLs carry a foreign canonical or stay unindexed. And when the variant pages you decided to keep are worth real visibility, give them proper product schema markup so the right URL is the one showing a price in the results.