SEO / Technical · 06 min read · Updated Sep 2026
Faceted navigation SEO: filters that publish a site you did not plan
Filters help people narrow a catalogue. Indexed combinations of those filters publish thousands of near-empty pages.
Index a facet only when it is a page a customer would search for. Leave the other combinations out of the index.
Separate useful facets from combinations
A single facet with real demand — a brand, a district, a material — can deserve a page if the copy and the inventory are unique. Every combination of colour, size, sort order, and availability usually does not.
Look at the crawl. If filter URLs outnumber products, the navigation is the site search has to understand.
Decide indexation facet by facet
Index the facet pages that match a query and have unique introductory content plus a stable inventory. Noindex or block the rest, including sort orders and multi-select combinations.
Use a clean path for the indexable facet and parameters for the views that should not be indexed. Then keep those parameters out of the sitemap and out of followable links where possible.
- Index one-facet pages only when demand and content are real.
- Keep sort and multi-select combinations out of the index.
- Canonicalise back only when the filtered view is the same list.
Watch crawl stats after the rules change
The goal is fewer fetched filter URLs and stable indexing of the facet pages you kept. If a rule blocks CSS or the product pages themselves, roll that pattern back.
Editors will ask for a new filter after launch. Add it to the same decision list before the template links it sitewide.
Questions
Should all filter URLs be canonicalised to the category?
Only when they show the same set of items. A facet with a genuinely different set can be its own page. A sort order should not be.
Is robots.txt enough to control facets?
It can stop the crawl of a parameter space. It will not noindex URLs that are already indexed. Pair the disallow with a plan for URLs already in the index.
Continue reading
- URL parameter indexing: tracking tags that become extra pages
- FAQ schema: markup for questions the page actually answers
- A practical content architecture for search intent
This note is part of the Hong Kong field library from My Search / Visibility, built with Itehk.