Most sites split long lists across pages: a blog archive, a product category, a services directory, a wall of reviews. Visitors move through those lists without thinking about it. Googlebot does not move through them the same way, and when the pagination is built the wrong way, everything past the first page can quietly drop out of Google’s reach. The pages still exist and still load in a browser. Google just never finds them, or finds them and folds them into page one.
How Google actually moves through a paginated list
Google’s documentation on pagination, incremental page loading and their impact on Search is specific about the mechanism. When it crawls a site to find pages, Google generally follows URLs found in the href attribute of <a> elements. Its crawlers do not click buttons, and they generally do not trigger JavaScript functions that need a user action to update the page. Google’s separate link best practices page says the same thing from the other direction: Google can reliably extract a URL only from an anchor element with an href, and not from elements that act as links through script events alone.
Put those two statements together and the rule for pagination is simple. A paginated list stays visible to Google when every page in the sequence has its own URL and the page before it links to that URL with a normal <a href>. Anything that reaches page 2 only through a click handler, a scroll event or a fragment is, for crawling purposes, a dead end.
Five ways pagination hides content from Google
These are the patterns we run into most often in site audits. Several of them are defaults in popular themes and page builders, so a site can have them without anyone having chosen them.
- Load more buttons with no fallback link: the button fetches the next batch with JavaScript, there is no
hrefto the next page anywhere in the HTML, and Googlebot, which does not click, only ever sees the first batch. - Infinite scroll with no paginated URLs underneath: new items append as the user scrolls, but there is no addressable URL for the second or tenth set of results, so there is nothing for Google to request or index.
- Page numbers in a URL fragment: links like
/blog#page=2look like separate pages, but Google ignores everything after the#, and its documentation warns it may not follow such a link at all because it thinks it already has that page. - Every page canonicalized to page one: a plugin or template sets
rel="canonical"on pages 2, 3 and 4 back to page 1, which tells Google the deeper pages are duplicates. Google’s guidance says the opposite: give each page in the sequence its own canonical URL. - Relying on rel=”next” and rel=”prev”: these link tags are still output by many themes and plugins, but Google states it no longer uses them. They are harmless, and other search engines may still read them, but they do nothing to help Google discover page 2.
The setup Google documents
The fix is not complicated, and it is the same whether the list is posts, products or locations. Google’s guidance comes down to three things.
Give each page a unique URL. A query parameter such as ?page=2 or a path such as /page/2/ both work. Google treats each URL in a paginated sequence as a separate page, which is what you want: each one is a crawlable path to the items listed on it.
Link the pages in order. Each page should link to the next one with a plain <a href>. Google also suggests linking every page in the collection back to page one, as a hint that the first page is the better landing page for the collection as a whole.
Let each page canonicalize to itself. Page 3 should carry a canonical pointing at page 3. If you want to understand why pointing everything at page one backfires, our earlier post on canonical tags covers how Google treats a canonical as a strong hint about duplicates, and Google’s own canonicalization documentation explains the signals it weighs.
One detail surprises people. Google’s documentation says paginated pages do not need distinct titles and meta descriptions; you can use the same title and description across the whole sequence, because Google tries to recognize pages in a sequence and index them accordingly. That is an exception to the normal advice to give every page a unique title, and it saves a lot of template work.
The paragraph worth keeping
Google stopped using rel="next" and rel="prev" to understand paginated content, and its current pagination documentation, last updated on 10 December 2025, says so directly. What Google relies on instead is three plain signals: a unique URL for every page in the sequence, a crawlable <a href> link from each page to the next, and a self-referencing canonical on every page rather than a canonical pointing back to page one. A “load more” button or infinite scroll can sit on top of that structure for visitors, but if the paginated URLs and links are not there underneath, Googlebot sees only the first batch of results.
Load more and infinite scroll can still work
None of this means abandoning load more or infinite scroll. The user-facing pattern and the crawlable structure are separate decisions.
The approach we use on web design builds is progressive. The server renders a normal paginated sequence with real links. JavaScript then upgrades the experience: the load more button intercepts the click and appends results, and the browser address updates to the matching paginated URL with the History API, so a visitor who shares or reloads lands on the right set. If the script fails, or if the visitor is Googlebot, the plain links still lead to every page.
Filters and sort orders are a different problem
Paging through a list should be crawlable. Re-sorting it by price or filtering it by size usually should not be indexed, because every variation is the same items in a different order. Google recommends keeping those variations out with a noindex robots meta tag, or discouraging crawling of those URL patterns in robots.txt; our recent post on faceted navigation covers that side. Make ?page=2 easy to reach, and make ?sort=price&page=2 something Google does not need to see.
A check you can run today
Open the longest list on your site, whether that is the blog archive, a category or a reviews page, and scroll to the bottom. Then check what Google can follow, not what a person can click.
- Open the browser’s developer console on page one and paste:
[...document.querySelectorAll('a[href]')].map(a => a.href).filter(h => /page[=\/]2/.test(h)). If the result is an empty array, there is no crawlable link to page 2 in the rendered page. - Load page 2 directly, using whatever URL pattern your site uses, and view the source. Find the
rel="canonical"line. It should point at page 2, not page 1. - Look at the next-page link itself. If its
hrefcontains a#, or there is nohrefat all and only a button, Google is not following it. - In Search Console, run URL Inspection on page 2 or 3 of the sequence. If it reports that the URL is unknown to Google, the sequence is not being discovered.
A sitemap that lists every item URL is a useful backstop, and Google suggests one for this exact situation. It does not replace the links, though. A sitemap tells Google a URL exists; the links between pages are what show how the collection fits together.
What we see in real audits
The pattern we encounter most is not a site that decided against crawlable pagination. It is a site that had it and lost it in a redesign. The old theme printed numbered page links; the new one ships with an attractive load more button that calls a JavaScript endpoint and prints no link at all. Nothing looks broken. The archive still works perfectly for visitors, and the only signal is a slow fade in impressions for older content in Search Console, often months after launch. That is why we check pagination on every redesign before it goes live, alongside redirects and canonicals, as part of our SEO services.
The same structure matters for AI-driven search features. Google says pages appearing in its AI features are drawn from the same index as normal results, with no special markup required, so a page Google cannot crawl is a page no AI answer will cite either. If you are working on answer engine optimization, crawlable access to your full archive comes before anything else.
The short version
Give every page in a list its own URL, link to it with a real anchor, and let it canonicalize to itself. Build load more or infinite scroll on top of that structure, never in place of it. Then run the console check above on your longest list.

