Updated
Pagination SEO means making page 2, 3 and onward of a list crawlable, so Google can reach every item on it, without letting those pages compete with page 1. Give each page its own URL such as ?page=2, link the pages with normal <a href> links, and give each page a canonical pointing to itself. Google no longer uses rel="next" and rel="prev".
Most guides to pagination and faceted navigation are written for online shops. Directories have the same problem with different nouns: a city page with 140 plumbers on it is a paginated list, and “open now”, “top rated” and “emergency service” are facets. This guide covers both, for listing sites.
What Google says about paginated lists
Google’s pagination and incremental loading guide is short and specific:
- One URL per page. Use something like
?page=n. Don’t use fragments (#page=2): “Google ignores fragment identifiers.” - Link the pages in order with
<a href>links from each page to the next, and Google suggests a link back to the first page. - Canonical to itself. “Don’t use the first page of a paginated sequence as the canonical page.” Page 2 has different listings from page 1, so it isn’t a duplicate of it.
- Titles can repeat. Paginated pages can share titles and descriptions, an exception to the usual advice.
- rel=next/prev: “Google no longer uses these tags, although these links may still be used by other search engines.” Leaving them in does no harm.
A paginated city page, done that way:
<!-- /plumbers/austin-tx?page=2 -->
<title>Plumbers in Austin, TX - Page 2</title>
<link rel="canonical" href="https://example.com/plumbers/austin-tx?page=2">
<nav aria-label="Pages">
<a href="/plumbers/austin-tx">1</a>
<a href="/plumbers/austin-tx?page=3">3</a>
<a href="/plumbers/austin-tx?page=3">Next</a>
</nav>Infinite scroll and “load more” buttons
Google’s crawlers don’t click buttons or scroll, so listings that only appear after “Load more” or scrolling may never be found. Keep the infinite scroll for visitors if you like it, but back it with real paginated URLs that the page links to, so each batch of listings has an address Google can fetch.
Paginated city and category pages on a directory
- Keep the order stable. If the sort order reshuffles on every request, the listings on page 3 change each time Google crawls it, and some listings are never seen on any page. Sort by something steady, such as completeness or name, with featured listings pinned first.
- Don’t make listings depend on page 30. A listing reachable only through deep pagination is weakly linked. Add other paths: nearby businesses on each listing, neighborhood or sub-area pages, “recently added” blocks.
- Return 404 past the last page.
/plumbers/austin-tx?page=99on a three-page list should not return page 1 or an empty page with a 200 status. - Don’t noindex page 2 and later as a blanket rule. They usually get no impressions, and that’s fine; their job is to get the listings on them crawled.
- Page size is a choice, not an SEO rule. Twenty to fifty listings a page is common. Fewer, longer pages mean fewer URLs to crawl.
Faceted navigation SEO for listing sites
Faceted navigation is the set of filters that narrow a list: category, city, rating, price, open now, features, sort order. Each filter, and each combination, can create a new URL. Google’s faceted navigation documentation names two problems that come with that:
- Overcrawling: the filter URLs “seem to be novel”, so crawlers fetch many useless URLs before recognizing them for what they are.
- Slower discovery: “if crawling is spent on useless URLs, the crawlers have less time to spend on new, useful URLs.”
The numbers grow fast. A directory with 20 categories and 200 cities has 4,000 category-in-city pages. Add five sort orders and ten pages each, and that’s 200,000 crawlable URLs, before a single filter is combined with another.
Google gives two routes. If you don’t need the filtered URLs in search, prevent crawling: robots.txt rules, or filters that change the page through URL fragments, which have “no impact on crawling (positive or negative)”. If you do need some of them indexed, Google’s best practices are:
- Use the standard
&separator between parameters. - Keep filters in a consistent order, with no duplicate filters, so one combination has one URL.
- Return a 404 when a filter combination has no results, and don’t redirect it to a generic not-found page.
- Use canonicals to point variants at the main version. Google notes this reduces crawling of those variants only “over time”.
- If you use
rel="nofollow"on filter links, it only works when every link to that URL has it.
Which filters deserve an indexable URL
The test is whether people search for that exact combination and whether you have enough listings to fill the page.
| Filter or combination | Indexable? | How |
|---|---|---|
| Category × city | Yes | A clean path such as /plumbers/austin-tx, in the sitemap, with enough listings |
| Category × city × a searched service | Sometimes | Only when people search it (“emergency plumbers austin”) and you have the listings; give it a clean URL like a category |
| Rating, price, “open now”, distance | No | Parameters, canonical to the clean page or blocked from crawling |
| Sort order, map view, page size | No | Never useful in search; block crawling |
Site search results (?q=) | No | Google’s URL guidelines suggest blocking dynamic search URLs in robots.txt |
| Two or more filters combined | No | The combinations multiply fastest; keep them out of crawling |
| Combination with zero results | No | Return 404, or don’t link to it at all |
Don’t link to empty options in the first place. Hiding a filter that would return nothing is better for visitors and leaves no empty URL for Google to find.
Crawl traps, and the robots.txt order of operations
A crawl trap is a set of links that generates URLs without end: sort and filter links on every page of every list, calendars that link to the next month forever, session IDs in URLs. Block the patterns you never want crawled:
User-agent: *
# Search results and sort orders: never useful in Google
Disallow: /search
Disallow: /*?*sort=
Disallow: /*?*view=mapOne order matters. If those URLs are already indexed, blocking them first freezes them in place: Google can’t crawl them, so it never sees the canonical or noindex you added, and Search Console may report them as “Indexed, though blocked by robots.txt”. Add the canonical or noindex first, wait until the URLs drop out of the index, then add the robots.txt rule. Blocked by robots.txt covers both statuses.
Common pagination mistakes on directories
| Mistake | What happens | Fix |
|---|---|---|
| Every page’s canonical points to page 1 | You tell Google that page 3 is a copy of page 1, though it lists different businesses. Google’s guide says not to do this. | Canonical each page to itself |
Page links are buttons with no href | Google doesn’t click, so it never reaches page 2 | Real <a href> links |
Page numbers in a fragment (#page=2) | Google ignores fragments, so every page looks like page 1 | A parameter or path, such as ?page=2 |
Page 1 reachable as both /plumbers/austin-tx and ?page=1 | Two URLs for one page | Link “1” to the clean URL; redirect or canonical the ?page=1 version to it |
| Out-of-range pages return 200 | Empty pages that Search Console may report as soft 404 | Return 404 past the last page |
Sort order carried into page links (?sort=rating&page=2) | Every sort order multiplies every page | Paginate the default order only; keep sorted views out of crawling |
Sitemaps don’t need page 2 and later. List the clean category-in-city URLs and the listings themselves; the listing sitemap is the safety net that gets every business found even if a list is long.
How to see which paginated and filter URLs Google shows
- Performance report, Pages tab, regex filter. Add a Page filter, choose Custom (regex), and enter
[?&]page=\d+(or/page/\d+) for paginated pages, or[?&](sort|rating|open)=for your filters. That shows how many of those URLs earn impressions, and for which searches. - Compare page 1 and page 2 for the same searches. If page 2 of a city list shows for “plumbers austin” as often as page 1, they compete. Keyword cannibalization covers finding those searches.
- Page indexing report. Parameter URLs tend to collect under Duplicate without user-selected canonical, Alternate page with proper canonical tag and Crawled - currently not indexed. The first is worth fixing; the second usually means your canonicals work.
- Crawl stats report (Settings). Its example URLs show whether Googlebot spends its time on filter and sort URLs instead of listings.
How DirectoryLens shows it
When you set up page types, Propose from URLs spots pagination and parameter URLs and gives them their own types, so their numbers stay out of your real page types. Then:
- The Thin and unused pages report lists every parameter combination Search Console reports, with how many URLs Google showed in 90 days, and counts page-2+ URLs and how many of them earn impressions.
- The Competing pages report flags searches where “page 2+ of a list competes with page 1” and where “a parameter (facet) URL competes with a clean URL”.
- Its recommendations follow the order above: parameter URLs with no impressions get “canonicalize or noindex, don’t block yet”; pagination gets “keep crawlable”.
For the wider picture of thin and empty pages, see index bloat and zombie pages. For listing sites that work like marketplaces, see marketplace SEO.