• Paginated pages should be crawlable, self-canonical and linked with normal anchor links.
• Google no longer uses rel=next and rel=prev.
• Infinite scroll and load-more buttons need real paginated URLs behind them.
• Decide which filter combinations deserve indexing and give those clean, static URLs.
• Block or canonicalise the rest, and never link to combinations that return zero results.

Category, listing and archive pages rarely fit on one screen. Pagination and filters solve that for users. For search engines they create a question: which of these thousands of URLs are worth crawling and indexing?

Pagination

Pagination exists so that crawlers can reach every product or article. That is its main SEO job. Page two onwards rarely ranks, but the items linked from it need to be discovered.

Best practice

• Give each page a unique URL such as ?page=2 or /page/2/.
• Self-referencing canonicals. Page two canonicalises to page two. Pointing everything at page one tells Google to ignore the items on deeper pages.
• Keep paginated pages indexable. Noindexing them causes Google to crawl them less and eventually stop following their links.
• Use plain HTML links with an href for next, previous and page numbers.
• Do not rely on rel=next and rel=prev. Google confirmed in 2019 that it had not used them for years. They do no harm and other systems may still read them.
• Reduce depth. Show more items per page and link to several pages at once (1, 2, 3 ... 10, 20) so nothing sits fifty clicks deep.

Infinite scroll and load more

Crawlers do not scroll or click. If more items load only through those actions, they are invisible. Keep paginated URLs underneath, update the address with the History API as the user scrolls, and include ordinary links to the next page in the HTML.

View-all pages

A single page listing everything can work for small sets, provided it loads quickly. For large catalogues it is usually too heavy.

Faceted navigation

Facets let users filter by size, colour, brand, price and so on. Each combination can produce a new URL, and the numbers multiply quickly: ten filters with ten values each produce more URLs than you will ever have products. Google has called faceted navigation the most common source of overcrawling complaints.

Step 1: decide what deserves to be indexed

Use keyword research. People search for "red running shoes" and "Nike running shoes", so those filtered pages have value. Nobody searches for "running shoes size 9 under £80 sorted by newest". Indexable facets should get clean, static URLs, unique titles and headings, a self-canonical and internal links.

Step 2: control everything else

• Robots.txt disallow is the most effective way to stop crawl waste on low-value combinations. It prevents crawling, so use it for spaces you never want explored.
• Canonical to the unfiltered category consolidates signals but the URLs are still crawled, and Google may ignore the hint if the content differs a lot.
• Noindex keeps pages out of the index but, again, they are still crawled.
• URL fragments or client-side filtering avoid creating crawlable URLs at all. Good for filters with no search value.
• Nofollow on filter links is a weak hint and not a reliable control on its own.

These can be combined by filter type, but not on the same URL in ways that conflict. A robots.txt block hides any canonical or noindex on that URL.

Hygiene rules

• Enforce a single parameter order.
• Return a 404 for filter combinations with no results, and do not link to them.
• Limit how many filters can be combined in a crawlable URL.
• Keep sort order out of the index entirely.

Common mistakes

• Canonicalising paginated pages to page one.
• Filters that create crawl traps of millions of URLs.
• Valuable filtered pages hidden behind JavaScript with no crawlable URL.
• Identical titles and headings on every paginated or filtered page.
• Blocking in robots.txt after the URLs are already indexed, which freezes them in place.

How to test

• Crawl with and without parameter URLs and compare totals.
• Check log files for the share of bot requests going to parameter URLs.
• Review Search Console's Page indexing report for large counts of "Crawled, currently not indexed" and "Duplicate without user-selected canonical".

The GEO angle

AI shopping and research assistants favour pages that answer a specific need. A well-built indexable facet page such as "waterproof trail running shoes for wide feet" is exactly the sort of specific result that query fan-out goes looking for, so the facet strategy that suits long-tail SEO suits AI search too.

Add your title here

This is a paragraph. Writing in paragraphs lets visitors find what they are looking for quickly and easily. Make sure the title suits the content of this text.