Pagination in 2026: rel=next Is Gone, Now What

Google stopped using the markup everyone implemented, which turned out not to matter. What does matter is whether anything on page forty is reachable at all, and most archives fail that test.

For years the correct answer to pagination was to add markup declaring which page came next and which came before. Then Google announced it had not been using that markup for indexing for some time, and had neglected to mention it.

A lot of people took that as a crisis. It was not. The markup was never doing much, and its removal exposed the thing that was actually broken, which is that on most large archives the items on later pages are effectively unreachable.

Should you remove the markup?

No. It is valid HTML, it costs nothing, and it is still meaningful to browsers for prefetching and to other systems that read page relationships. Leave it where it exists.

What changed is that you cannot rely on it to communicate anything to Google. Each page in a series is treated as its own page, discovered through the links on it, and evaluated on its own content.

That shifts the whole problem from declaring relationships to making sure the pages are crawlable and the links on them work.

The three rules that replaced it

Each paginated page is self-canonical. Page two is its own page with its own items. Canonicalising it to page one declares it a duplicate, which tells a crawler that everything linked only from page two is reachable from somewhere redundant. This is the most damaging pagination mistake and we covered it among the five canonical faults in our piece on canonical tags.

Each page must carry real links. Actual anchor elements with href attributes, present in the HTML without needing JavaScript. A next button that works through a click handler is not a link, and nothing follows it.

Each page must be crawlable. Not blocked, not behind a parameter rule that strips it, not requiring a session.

Get those three right and pagination works. Everything else is detail.

Should pages two onwards be indexed?

This is the question people spend the most time on and it matters the least.

Paginated archive pages rarely rank for anything, because they are lists of titles with no original content. Whether they sit in the index is close to irrelevant.

What is not irrelevant is whether they are crawled, because they carry the links to your actual content. So if you decide to keep them out of the index, the directive must still permit following.

Be aware of the long-term behaviour. A page kept out of the index indefinitely tends to be crawled less over time, and the links on a rarely crawled page do less work. For that reason our default is to leave paginated pages indexable and simply accept that they will not rank. They are plumbing, and plumbing does not need to be hidden.

The decision framework between the available tools is in robots.txt, noindex or canonical.

URL design

Either a path segment or a query parameter works. Be consistent, and pick one.

Two things to avoid. Do not let page one exist at two addresses, both with and without a page indicator, which creates a duplicate pair at the most important URL in the series. And do not combine the page number with parameters that also control sorting or filtering unless you have thought about the combinations, because that is how an archive of forty pages becomes several thousand URLs.

The combinatorial version of this problem is the subject of faceted navigation and pagination.

The real problem: depth

Here is what the markup discussion always distracted from.

If your archive has eighty pages, the items on page sixty are sixty clicks deep from your main navigation. A crawler working through a site allocates effort by importance and reachability, and something sixty hops down a chain of near-identical list pages is as close to invisible as a live URL gets.

This is not fixed by markup, by canonicals or by indexing directives. It is fixed by making the content reachable another way.

  • Better taxonomy. If people can reach items through categories, topics or filters, the depth from the home page drops dramatically. Nobody should have to paginate to find something.
  • Fewer items per archive. Splitting one enormous archive into several meaningful ones is usually better than paginating one of them forever.
  • Internal links from real content. An item linked from three articles is reachable regardless of where it sits in a list.
  • Sitemaps as the backstop. They are a discovery mechanism, not a substitute for structure, and they ensure nothing is dependent on being paginated to. The structure is in sitemaps that scale.

If the honest answer is that a tenth of your content is only reachable by paginating thirty pages deep, that is a structural finding rather than a pagination setting, and it will explain more about your indexing than anything else you investigate. Crawl budget explained covers why.

How many items per page

A real decision with trade-offs in both directions, and most sites have never made it deliberately.

Fewer items per page means more pages, more URLs, greater depth, and a heavier crawl footprint for the same content.

More items per page means fewer pages and shallower depth, at the cost of a heavier page. A listing page with two hundred items is slow to render and unpleasant on a phone, which is where most Indian traffic arrives.

Our rule of thumb, and it is ours rather than a standard: enough items that the series stays under about ten pages, and few enough that the page stays light. For most archives that lands somewhere between twenty and fifty. If that arithmetic does not work for your volume, the answer is more categories rather than a different number.

View all pages

A single page containing every item, offered alongside the paginated series. It was once a recommended pattern and it still has narrow uses.

It works when the complete set is genuinely modest, a few dozen items, and when the page stays fast. It fails on anything large, where you have simply moved the problem into a page that takes eight seconds to render.

If you offer one, make it the canonical version and have the paginated pages reference it. Do not run both as independent indexable series, which is a duplication problem you created for yourself, as in handling duplicate content.

Load more buttons and infinite scroll

Both replace pagination links with a script, and both break discovery unless real paginated URLs exist underneath.

The workable pattern is to keep genuine paginated URLs that function without JavaScript, and layer the button or the scroll behaviour on top for people. A crawler follows the links, a visitor never sees them.

The implementation detail is enough of a subject on its own that we deal with it separately.

What is a waste of effort

Removing the next and previous markup because Google stopped using it. It is harmless.

Trying to make page two rank. It is a list of titles.

Adding unique written content to each paginated page to make it substantial. Nobody reads it, it is written for a crawler, and it does not change the outcome.

Blocking paginated pages in robots.txt. You have now prevented the crawler from following the links on them, which is the only thing they were for.

Optimising pagination while leaving an archive eighty pages deep. You are tuning the symptom.

A sensible order of work

  1. Crawl your site and find your deepest content. Count the clicks from the home page.
  2. Confirm every paginated page is self-canonical.
  3. Confirm the next and previous controls are real links in the HTML.
  4. Confirm paginated pages are not blocked and are permitted to be followed.
  5. Check that page one does not exist at two addresses.
  6. Count your pages per archive. If any series runs past about ten pages, add categories rather than adjusting the page size.
  7. Make sure every item appears in a sitemap regardless of where it sits in a series.
  8. Add internal links from real content to the items that currently depend on pagination.

If the goal is getting new items seen quickly rather than fixing an archive, getting new pages discovered and indexed quickly is the more direct answer.

If you want a crawl run across your site to find what is only reachable by paginating, get in touch.