Pagination, load more and infinite scroll: what Google crawls
Short answer: Google finds the products or posts beyond your first screen by following ordinary links, and its crawler doesn't click buttons or scroll. So whatever your visitors see, page 2, 3 and onward need their own URLs (like ?page=2) linked with real <a href> tags. Load more and infinite scroll are fine for visitors as long as those paginated URLs exist underneath. Give each page its own canonical, not page one's.
Most small shops never think about this until a product launched last spring still doesn't show up in search. Often the reason is simple: the only route to it is a "Load more" button or a scroll that Google never triggers.
What Google actually does on a list page
Google's pagination and incremental loading guide explains the mechanics. When crawling, Google generally follows URLs in the href attribute of <a> elements. Its crawlers don't "click" buttons and generally don't trigger JavaScript functions that need a user action to update the page.
That one fact explains most problems. A category page that shows 24 products and loads the rest by JavaScript on click or scroll may be showing Google only those 24.
The guide also suggests a sitemap or a Google Merchant Center feed to help Google find all your products. Those help, but they don't replace crawlable links, which also tell Google how pages relate.
The three patterns, briefly
Google's page describes three UX patterns and their trade-offs for visitors:
- Pagination: "next", "previous" and page numbers. It shows people where they are and how much there is, but needs a new page load to see more.
- Load more: a button that extends the list. One page for everything, and the button can show the total, but it can't handle very large result sets.
- Infinite scroll: more items load as you scroll. Intuitive, but Google notes it can cause "scrolling fatigue" because the size is unclear, and it can't handle very large numbers either.
Pick the pattern for your visitors first. Then make sure the crawlable version exists behind it.
Good and risky setups
| Setup | Search risk | Why |
|---|---|---|
Numbered pages at ?page=2, ?page=3, each linked with <a href> |
Low | Each page has a unique URL Google can follow |
Load more button that also updates the URL to ?page=2 and has a linked fallback |
Low | Visitors get one page; crawlers get real URLs |
| Infinite scroll backed by component pages that work without JavaScript | Low | Matches Google's search-friendly recommendations |
| Load more button with no URLs behind it | High | Google doesn't click it, so later items may be missed |
| Infinite scroll with no paginated version | High | Items after the first batch may not be discoverable |
Page numbers in fragments (#page=2) |
High | Google ignores fragments and may not follow the link |
| Every page canonicalized to page 1 | High | Google's guide says not to do this; it hides later pages |
Sort and filter variations (?order=price) all indexable |
Medium | Duplicate lists; Google suggests noindex or robots.txt rules |
Relying on rel="next"/rel="prev" alone |
Medium | Google says it no longer uses them |
How to set up paginated URLs properly
Give every page its own URL
Google's guide is explicit: give each page a unique URL, for example with a ?page=n parameter, because pages in a paginated sequence are treated as separate pages. Don't use fragment identifiers for page numbers.
Link pages in sequence
Each page should link to the next with a normal <a href>. The guide also suggests linking every page in the collection back to the first page, which hints that page one is the best landing page.
Canonicals: self, not page one
Each paginated page gets its own canonical URL pointing to itself. Pointing page 2 and page 3 at page 1 tells Google they're duplicates, and the products only listed on those pages lose their path in.
Titles can repeat
Normally every page should have a distinct title. For a paginated sequence, Google says you can use the same title and description on each page; it tries to recognize sequences on its own.
Keep filters and sorts out of the index
If you support ?order=price or color and size filters, Google recommends blocking unwanted variants with a noindex robots meta tag or discouraging crawling of those patterns in robots.txt. Faceted navigation is a topic of its own; the short version is to decide which filter combinations deserve to be landing pages and keep the rest out.
Making infinite scroll search-friendly
Google's older but still-relevant Search Central blog post on infinite scroll sets out the approach:
- Split the infinite list into component pages that can be accessed when JavaScript is disabled.
- Make sure each item appears once in the series, with no overlap between pages.
- Size each page so a visitor arriving directly can find what they came for without endless scrolling, with reasonable load time.
- Update the address bar as people scroll, using
pushStateand/orreplaceState, so the URL matches the part of the list they're viewing. - Test that out-of-range pages (page 99 of a 12-page list) return a 404.
Note that the post dates from 2014 and mentions rel="next" and rel="prev", which Google's current guide says it no longer uses. The component-page idea is the part that still matters.
A hypothetical check for your own shop
Imagine a hypothetical online tea shop with 140 products in one category, shown 24 at a time with a Load more button.
- Open the category in a private window and disable JavaScript. Can you reach products 25 to 140 through links?
- View the page source and search for
?page=2. Is it in an<a href>? - Check the canonical tag on page 2 if it exists. Does it point to itself?
- In Search Console, inspect one product that only appears after "Load more". Is it indexed, and how was it discovered?
If steps 1 and 2 fail, add a paginated fallback below the button; most ecommerce platforms and themes can do this without changing the visible design.
FAQ
Is "View all" on one page a good alternative?
For a modest list, yes, if it loads quickly. Google's own pros and cons note that single-page patterns can't handle very large result sets, so it's a poor fit for big catalogs.
Should paginated pages be noindexed?
Not by default. If page 2 is the only place some products are linked from, noindexing it can make those products harder to find. Noindex the sort and filter variants instead.
Do I still need a sitemap?
Yes. Google recommends a sitemap or a Merchant Center feed alongside crawlable links, not instead of them.
Does this apply to blog archives too?
Exactly the same rules. Older posts reached only through infinite scroll are at the same risk as products.
A category that hides half its products behind a button is a fixable problem, and we can check how Google sees yours before you redesign anything.