Pagination SEO

Long definition

A paginated series presents one logical list across multiple URLs: /category, /category?page=2, /category?page=3. Each page is a real, distinct URL with different items. The SEO question is how to signal the relationship to Google and which page should rank.

Google's current guidance, after the March 2019 deprecation of rel=prev/next, is straightforward:

• Self-canonical each page. <link rel="canonical" href="/category?page=2"> on page 2. Do not canonical page 2+ back to page 1 — that drops items 21+ from the index entirely.
• No rel=prev/next needed. Google ignores them. Bing and accessibility tools still respect them, so leaving them in is harmless but optional.
• Crawlable links between pages. Standard <a href> to neighbor pages and a few jump targets (first, last, ±2). Internal links carry PageRank between pages; without them, deep pages become orphaned.
• Unique title and description per page. "Category — Page 2" beats identical metadata. Don't manually rewrite each one — most CMSs append the page number automatically.
• noindex deep, low-value pages selectively. A 200-page archive of a small blog rarely deserves indexing past page 5. noindex,follow on deep pages preserves crawlability without polluting the index.

The view-all alternative: render the entire list on one URL (/category?view=all) and canonicalize paginated pages to it. Works only when the full list is small enough to load fast — under a few hundred items as a rough ceiling. Above that, view-all becomes a Core Web Vitals problem; stick with pagination.

The faceted-navigation overlap matters: a paginated filter (/category?color=red&page=2) compounds the URL space. Decide which combinations to index and use canonicals or noindex rules to keep the rest from diluting signals. Pagination on top of faceting is where most large e-commerce indexability problems start.

Common misconceptions

• "Page 2+ should canonical to page 1." It shouldn't. Google has been explicit: a non-self canonical on page 2 tells Google "page 2 is a duplicate of page 1", and the items that exist only on page 2 disappear from the index.
• "rel=prev/next is still required." Google deprecated it for indexing in 2019. Leaving it in for Bing and accessibility is fine; relying on it for Google ranking is not.
• "Every page in the series should rank." Usually no. Page 1 should rank for the category query. Pages 2+ exist to expose deeper items to crawlers and to users browsing the full list, not to compete in SERPs.
• "View-all always beats pagination." It only beats pagination when the full list loads quickly. A 5,000-item view-all destroys LCP and can be slower for users than a paginated alternative. Test before defaulting.