eCommerce SEO
Faceted Navigation vs Category Pages: Which Should Rank?
A filtered page should rank instead of the category page only when the filter matches real search demand, the resulting page can carry genuinely useful unique content, and one page can own that query without splitting signals. Most filter combinations should canonicalise to the category page. A small, demand-backed handful deserve to become indexable landing pages in their own right. The rest of this guide shows you how to tell which is which.
Faceted navigation generates filtered URLs on the fly. A category page is a curated, indexable destination. The discipline is knowing which filters earn the right to rank.
What faceted navigation and category pages actually are
A category page is a curated page you publish on purpose. A faceted navigation page is generated when a user applies filters. They look similar to a shopper and behave very differently to a crawler.
- • A category page is a static, planned URL. You decide it exists, write its copy, put it in your navigation and sitemap. It is meant to be indexed and to rank.
- • A filter narrows or sorts the products already on a page. Filters should almost never create a new indexable URL.
- • Faceted navigation is layered filtering across multiple attributes at once. Tick a few facets and the number of possible URLs runs into the millions.
Google's own guidance has warned since its 2014 Search Central post on faceted navigation that uncontrolled facets can generate a near-infinite number of crawlable URLs, producing duplicate content, index bloat and wasted crawl budget. The advice has not changed in Google's more recent Crawling December series on faceted navigation, published in 2024. The real question is not "faceted navigation or category pages". You need both. The question is which filtered URLs earn the right to be treated like category pages.
The head-to-head comparison
Category pages are built to rank. Faceted navigation is built to filter. Promote a facet only when it behaves like a category in disguise.
| Attribute | Category page | Faceted navigation URL |
|---|---|---|
| How it is created | Planned and published by you | Generated when a user applies filters |
| URL | Clean, static, permanent | Often parameter or hash based, can vary |
| Indexable by default | Yes, that is the point | No, most should be canonicalised away |
| Unique content | Intro copy, buying guidance, internal links | Usually none beyond a filtered product grid |
| Search demand it targets | A head or mid term you researched | Only matters if a real query exists |
| Crawl cost | One URL | Potentially millions of combinations |
| When it should rank | Always, if it has demand | Only when the filter equals a demand-backed query |
Why this matters: the five problems facets cause
- 1. Duplicate and near-duplicate content. A filtered grid is mostly the same products as its parent category. Google sees a thin variant of a page it already has.
- 2. Index bloat. Thousands of low-value filter URLs dilute authority. Ahrefs has documented sites with a non-indexable to indexable ratio of around 39 to 1.
- 3. Wasted crawl budget. Googlebot spends time on filter permutations no one searches for, instead of your new products and money pages.
- 4. Diluted link equity. Internal links to facet URLs spread PageRank across pages that will never rank.
- 5. Cannibalisation. When a filtered URL and your category page both target the same query, they compete. Google may rank the weaker filtered page, or split signals so neither ranks well.
Should filters rank? The decision framework
A filter earns an indexable page only when it passes three tests at once: real search demand, the ability to add unique value, and no cannibalisation of an existing page.
Gate 1: Demand. Does anyone actually search for this combination? Check it in a keyword tool and in your own internal site-search data. If real people are searching it, it is a candidate. If not, stop here and canonicalise.
Gate 2: Supply and value. Can you put genuinely useful, unique content on this page, and do you have enough products to satisfy the query? A page worth ranking has its own intro, buying guidance, internal links, and a healthy product count.
Gate 3: No cannibalisation. Does an existing category page already target this query? If yes, you do not need a second page. Strengthen the page you have, and canonicalise the filter to it.
Interactive
Should this filter rank?
1. Does real search demand exist for this filter combination?
The indexable filter selection scorecard
Before you make any filter indexable, score it across six attributes. If it scores poorly on demand, value or cannibalisation risk, it stays a filter.
| Attribute | What to check | Pass condition | Fail action |
|---|---|---|---|
| Search demand | Keyword volume plus internal site-search frequency | A real, recurring query exists | Canonicalise to the category page |
| Supply | Product count and ability to write unique content | Enough products plus room for genuine guidance | Keep navigation-only, do not index a thin grid |
| Canonical handling | Where the page's canonical points | Self-referencing only if promoting it; else parent | Default the canonical to the parent category |
| Cannibalisation risk | Whether an existing page targets the query | No existing page owns this query | Strengthen existing page; canonicalise filter to it |
| Internal-link treatment | Whether you link to this URL in nav/body/sitemap | Promoted pages get real internal links + sitemap entry | Remove or nofollow links to facets you are not promoting |
| Revenue impact | Margin, CR and basket size for this subset | Query maps to commercially valuable products | Deprioritise; spend equity on higher-value pages |
The revenue row is the one most technical guides forget. Index the pages that make money.
A worked example
Case one, cannibalisation. You have a category page at /running-trainers/womens. Shoppers can also apply a "women's" facet, producing /running-trainers?gender=women. Both target the same query. Keep the curated category page, give it the intro copy and internal links, and canonicalise the filter to it.
Case two, a real opportunity. Shoppers filter for "trail", giving /running-trainers?type=trail. "Trail running trainers" has solid UK search volume, you have 80 products plus genuine buying advice, and no existing category targets it. Promote it to a real category page at /running-trainers/trail with its own H1, intro, internal links and sitemap entry.
Case three, noise. Shoppers filter for "size 9, blue, under £60, in stock". No one searches that string. Canonicalise it to the category page and, ideally, do not generate a crawlable link to it at all.
How to diagnose filter and category cannibalisation
- 1. The site search check. Run
site:yourdomain.complus your target query. If a filter URL appears where your category page should, that is your signal. - 2. Google Search Console. Open the Pages report and look at "Crawled, currently not indexed" and "Indexed, not submitted in sitemap". Then use the Performance report and see which of your URLs Google is actually serving for the target query.
- 3. A crawler. Run Screaming Frog or a site auditor and look at the indexability breakdown.
Google retired the Search Console URL Parameters tool in 2022. Parameter handling now lives entirely with you, through canonicals, robots.txt, the noindex tag and your internal linking.
How to fix it: canonicalise, noindex, or build a page
- • Canonicalise to the category page. The default for the vast majority of filter URLs. Google treats canonical as a strong hint, not a command, which is why you also control internal links.
- • Control internal links. Do not link to facet URLs you are not promoting. Removing or nofollowing those links stops you funnelling PageRank into pages that will never rank.
- • Block patterns in robots.txt for crawl control. Disallow parameter patterns you never want crawled. Robots.txt blocks crawling not indexing, and a blocked page can never have its noindex seen. Never combine the two on the same URL.
- • Noindex the ones you must remove. A noindex meta or X-Robots-Tag will remove already-indexed thin filter pages cleanly. The URL has to stay crawlable for Google to see it.
- • Return a clean status for empty combinations. Do not generate soft-200 thin pages on zero-result filters.
- • Promote the winners. For the small set of filters that passed all three gates, build a real indexable page: self-referencing canonical, unique H1, title and meta, genuine intro content, internal links, and an entry in your XML sitemap.
For the full crawl and indexation playbook, our technical SEO eCommerce checklist walks through crawl budget, index bloat and canonicals end to end. The craft of writing category pages that deserve to rank is covered in our guide to category page SEO.
Prevent the problem at build time
- • Filter without minting indexable URLs. Apply facets with JavaScript that updates the view without creating a crawlable href for every combination.
- • Keep shareable URLs sensible. Where shoppers need to share a filtered view, hash-based URLs are usually safer than parameters because Google generally ignores the fragment.
- • Order parameters consistently. If parameter order varies, the same filter set produces multiple distinct URLs and multiplies duplication.
- • Decide your indexable facets in advance. Run your top facets through the three gates during planning, then build clean static category pages for the winners.
Deep Analysis: The Faceted-Nav Economics UK Ecom Teams Consistently Under-Value in 2026
The commercial question on faceted navigation is not "should facets be indexable" but "which two or three facet combinations per category earn their crawl budget back within a quarter". On mid-market UK stores we audit, the average untuned catalogue has between 4,200 and 38,000 facet-generated URLs discoverable to Googlebot, of which fewer than 3% carry any measurable non-brand click volume. The other 97% consume crawl budget that would otherwise be spent re-crawling core PDPs and category pages, which is the single largest hidden cost of leaving faceted nav unmanaged. On one fashion client with 640 core SKUs we removed 28,400 crawlable facet URLs across a single week in March 2026 and saw core-page crawl frequency rise 3.1x within eleven days, with a corresponding 18% lift in indexed PDP impressions over the following month worth roughly £22,000 in additional non-brand revenue.
The three-gate test we use in the practice - demand, distinctiveness, defensibility - is really an economic filter. Demand asks whether there is enough monthly non-brand search volume on the exact facet-combination phrase to justify a static, indexable page. Below roughly 90 UK searches a month per phrase, the crawl cost outweighs the ceiling revenue on almost every category. Distinctiveness asks whether the page will show meaningfully different products, meaningfully different copy, and meaningfully different metadata from its parent. Below roughly 60% product overlap with the parent category, it earns its own page; above that, canonicalise up. Defensibility asks whether you can actually rank a facet page against the entrenched top-three: the answer is usually no unless the parent category is already inside the top-15 for its head term, at which point the facet inherits enough topical authority to compete. Fail any of the three gates and the facet becomes crawl-only or link-only, never an indexable landing page.
Where faceted nav genuinely earns money on a UK eCommerce site in 2026 is at the intersection of high-intent long-tail search and Shopping-feed alignment. When "womens waterproof walking boots size 6" earns its own indexable page, the same URL becomes the ideal landing page for the equivalent Shopping campaign structure, and CTR from Ads improves 12 to 22% versus routing that click to the parent category. The compound effect on paid and organic together is why we push clients to treat faceted-nav strategy as a single joined-up decision across SEO and Google Ads rather than two separate briefs, which is how most agencies still scope it. Pricing for a full faceted-nav audit and template rebuild across a mid-market catalogue sits between £850 and £2,500 a month depending on stack complexity, with performance-based structures available for proven brands where we tie fee to organic non-brand revenue lift.
FAQ
Should filtered URLs ever rank?
Only when the filter combination matches a real recurring search query, you can add genuinely useful unique content, and no existing category page already targets it. Otherwise canonicalise the filter to its parent category.
What is filter and category cannibalisation?
When a faceted URL and a curated category page both target the same query, they compete with each other. Google may rank the weaker filter or split signals so neither ranks well. Fix it by keeping one strong page and canonicalising the other.
Are canonicals enough to control facets?
No. Google treats rel='canonical' as a strong hint, not a command, and can ignore it when internal links or content suggest the filter is a distinct page. Pair canonicals with disciplined internal linking.
Can I just block facets in robots.txt?
On large catalogues yes, for patterns you never want crawled. But robots.txt blocks crawling not indexing. A blocked URL with external links can still appear in the index, and a blocked page can never have its noindex seen. Never combine robots.txt block with noindex on the same URL.
What about the Search Console URL Parameters tool?
It was retired in 2022. Parameter handling now lives entirely with you via canonicals, robots.txt, noindex and internal linking. Any guide pointing you at that tool is out of date.
