Canonical Tags and Pagination

How to canonicalize a paginated series

A paginated series is a set of pages that splits a single collection of content into sequential chunks: page 1, page 2, page 3, and so on, of a category listing, a search results set, or a long article broken into parts. The correct canonicalization approach is simple and often gets over-thought: each page in the series should carry a self-referencing rel=canonical, meaning page 2 canonicalizes to page 2's own URL, page 3 canonicalizes to page 3's own URL, and so forth. There is no cross-page canonicalization within a normal paginated series.

This follows directly from what rel=canonical is for. It is a hint to Google about which URL is the preferred version among a set of duplicate or near-duplicate URLs. Pages 2, 3, and 4 of a paginated series are not duplicates of page 1 or of each other. They display different items. Treating them as duplicates by canonicalizing them elsewhere misrepresents what is actually on the page, and it is the single biggest mistake site owners make with pagination.

Each paginated URL should also remain indexable: no noindex tag, no robots.txt block, and a normal self-referencing canonical in the raw HTML head (or HTTP header, not JavaScript-only). Within the paginated pages, link individually to each item so Googlebot can discover and crawl them through normal on-page links, since that is how Google finds paginated content today.

Why rel=prev/next no longer matters to Google

For years, many SEOs treated the rel=prev and rel=next link elements as an important signal for helping Google understand pagination. On March 21, 2019, Google clarified that it no longer uses rel=prev/next for indexing purposes, and had not for years by that point. Googlebot discovers paginated pages through on-page links rather than through these attributes.

That does not mean the attributes are invalid or that you need to strip them out. They remain valid HTML, and they may still be used by other search engines or by assistive technology for accessibility purposes. If your CMS or theme already outputs rel=prev/next, there is no harm in leaving it in place. The point is simply that from Google's perspective, these attributes carry no indexing weight, so you should not build a pagination strategy around them, and you should not treat their absence as a problem worth fixing.

What actually matters for Google to understand and crawl a paginated series is straightforward crawlable links between pages, each page returning a normal 200 status, each page being indexable, and each page self-canonicalizing as described above.

The mistake of canonicalizing every page to page one

The most common and most damaging pagination mistake is canonicalizing every page in a series to page 1. This usually comes from a well-intentioned but mistaken instinct: since pages 2, 3, and 4 feel like variations of the same listing, someone assumes they are duplicate content and consolidates them onto the first page, the way you might consolidate a URL with tracking parameters onto its clean version.

This reasoning is wrong because pages 2 and beyond are not duplicates of page 1. They contain different items, different products, different articles, or different search results. Canonicalizing them to page 1 tells Google, in effect, that there is only one page in the series worth indexing. The practical result is that Google may not index the deeper pages at all, and it may ignore links and other signals that exist only on those deeper pages, since a page that is canonicalized elsewhere is treated as non-canonical content.

The damage compounds for any site where deeper pagination pages are the only place certain items are linked from. If product pages 40 through 80 in a large catalog are only discoverable via links on paginated page 3 and beyond, and those pages are all canonicalized to page 1, you have effectively told Google to disregard the very pages that provide the crawl path to that inventory. This is a common cause of large chunks of a catalog or archive failing to get indexed, and it traces directly back to a pagination canonical mistake.

The correct self-referencing approach

The fix is to let each paginated page canonicalize to itself. In the head of page 2 of a category listing, the tag should read exactly like this:

<link rel="canonical" href="https://example.com/category/widgets/page/2/" />

And on page 3, the canonical points to page 3's own URL, and so on for every page in the series. This is a self-referencing canonical, and it is the standard, recommended approach for pagination. As with any self-referencing canonical, the URL must be absolute, must exactly match the served protocol, host, casing, and trailing-slash convention, and should live in the raw HTML head or HTTP header rather than be injected only by JavaScript.

There is one legitimate exception. If a genuine view-all page exists, one that loads the entire collection on a single URL and performs acceptably for users, it can be appropriate to canonicalize the individual paginated pages to that view-all page instead of self-referencing. The caveat is that the view-all page has to actually be a good user experience: it needs to load and render properly with everything included, not be a thin stub or a page that times out under the weight of the full collection. If no such page exists, do not manufacture one purely to consolidate pagination; self-referencing canonicals on each paginated page remain the correct default.

How to audit pagination canonicals

Auditing pagination canonicals is a matter of checking a sample of paginated URLs across every paginated template on the site: category or archive listings, search results, and any multi-page articles. For each page beyond page 1, confirm that the rel=canonical in the raw HTML points to that page's own URL, not to page 1 and not to some other page in the series.

A site crawler that reports on canonical relationships is the fastest way to do this at scale, since it can flag every paginated URL whose canonical target differs from itself. Look specifically for a pattern where large numbers of URLs matching a pagination structure, such as those containing "/page/" or a page query parameter, all canonicalize to the same first-page URL; that pattern is the signature of the page-one mistake described above.

It is also worth spot-checking a few deeper pagination pages in Google Search Console's URL Inspection tool to see whether Google is respecting the self-referencing canonical or selecting a different canonical of its own. Since rel=canonical is a hint rather than a directive, Google can still override it based on other signals, so confirming what Google actually chose is a useful final step beyond just verifying the tag is correct in the HTML.

Frequently Asked Questions