Faceted navigation is useful for shoppers and risky for search. The job is to decide, facet by facet, which URLs earn a place in the index and which should be controlled.
Quick answer
Treat each filter as one of three things. A few facets match real search demand and deserve a proper, indexable landing page. Many create near-duplicate versions of a category and should point back to it with a canonical tag. The rest, such as sort orders, multi-select combinations and empty result pages, should be kept out of the index altogether. Decide this per facet, not for the whole site, and check the result in Search Console afterwards.
What faceted navigation does to a site
A category page with filters for size, colour and type can generate a new URL for every combination. Three filters with a handful of options each can produce hundreds of URLs that show almost the same products as the parent category.
That causes three problems.
- Duplication. Near-identical pages compete with the category page and with each other, and link signals get split between them.
- Index bloat. Google may index many filtered URLs that were never meant to rank, and crawl time is spent on them instead of on pages that matter.
- Wasted crawling. On large catalogues, a large share of Googlebot's visits can go to filter combinations. On a small site this is less of a concern than the duplication.
A real example
A UK plastic storage retailer I audited had 288 indexed URLs on a self-hosted Magento site. Roughly 130 of those were near-identical variants of existing category pages, produced by filtering by size, colour and type, and there were no canonical tags to tell Google which version mattered.
Canonical tags were added across the category and product pages as part of a five-week fix programme. At the 90-day mark the number of duplicate pages indexed had fallen from about 130 to 8. The full detail, including the other fixes made at the same time, is in the storage retailer case study.
On a catalogue that size, the main damage was duplication and split signals rather than crawl waste. On a catalogue with tens of thousands of products, the same pattern would put far more pressure on crawling as well.
Step 1: find out how many facet URLs exist
Start with evidence, not assumptions.
- Crawl the site with parameters and filtered URLs included, and count how many URLs come from facets compared with real categories and products
- Check the Pages report in Search Console for filtered URLs that are indexed, and for "Duplicate, Google chose different canonical than user" entries
- Look at the Crawl Stats report to see whether crawl requests are heavily weighted towards parameter URLs
- Search for your own site with a filter parameter in the query to see what Google has chosen to index
If you only find a handful of filter URLs, the fix is small. If you find thousands, the plan needs more care.
Step 2: sort every facet into three groups
Go through each filter and decide which group it belongs to.
Group 1: index it
A facet deserves its own indexable page when people actually search for it and the page can offer something distinct. A product type combined with a size, or a brand within a category, are typical examples. A colour filter on most categories is not.
Check search demand before deciding. If nobody searches for "red plastic storage boxes size 5", there is no reason to give that combination its own page.
Group 2: canonicalise it
Where a filter changes the products shown but does not deserve its own page, point the filtered URL to the main category with a canonical tag. This is the right default for most single filters. It tells Google which version to consolidate on while letting shoppers keep using the filter.
Group 3: keep it out
Some URLs should not be in search at all. Sort orders, items-per-page settings, multi-select combinations, and filter combinations that return no products belong here. Use a noindex tag, or keep them from being crawled in the first place.
Step 3: use the right control for each group
The controls do different jobs, and mixing them up is a common cause of faceted navigation problems.
- Canonical tag: a hint that a page is a duplicate of another. Google may ignore it if the pages are too different, so do not rely on it for pages that should stay out of the index.
- Noindex: tells Google not to index a page. It only works if Google can crawl the page and see the tag.
- robots.txt disallow: stops crawling, which saves crawl time, but Google cannot then see a noindex or canonical on the blocked page. Do not combine a disallow with a noindex and expect both to work.
- Internal links: a link to a filtered URL signals that it matters. Remove or reduce links to facets that should not rank.
For indexable facets, make them proper pages rather than parameter noise: a clean, stable URL, a unique title and heading, some useful copy, a self-referencing canonical, a place in the sitemap and links from the parent category.
Step 4: check it worked
After the changes go live, recheck after a few weeks.
- Is the count of indexed filter URLs falling in the Pages report?
- Are the facets you chose to index now indexed, and getting impressions?
- Has the crawl split in Crawl Stats moved away from parameter URLs?
- Have any category pages lost visibility they should have kept?
If any indexable facet is not being picked up, check that it is linked internally, included in the sitemap and not blocked or noindexed by mistake.
Common mistakes
- Indexing everything because "more pages means more traffic"
- Blocking everything and losing facets that match real demand
- Using robots.txt to hide pages that are already indexed, which stops Google seeing the instruction to remove them
- Relying on canonicals alone for pages that are not near-duplicates
- Ignoring internal links, so filtered URLs keep receiving the signals the controls are trying to remove
- Not measuring afterwards, so nobody knows whether the change helped
When to get support
If a category-heavy site has an unclear index count or filtered URLs showing up in search, a technical SEO review can map which facets to keep, canonicalise or block. A Visibility Review checks for duplicate URL patterns like this, and the SEO audits page explains the wider scope.
FAQ
Should filtered pages ever be indexed?
Yes, when a filter matches real search demand and the page can offer unique value, such as a product type with a size. Most single filters, and almost all combinations, should not be.
Is canonical or noindex better for facets?
Use canonical for filtered pages that are near-duplicates of a category, and noindex for URLs that should not be in search at all. Use each for the job it does.
Does faceted navigation waste crawl budget?
It can, especially on large catalogues where filter combinations multiply. On smaller sites the more common damage is duplication and split signals. Check Crawl Stats before assuming crawl budget is the problem.
Should I block facets in robots.txt?
Only for patterns that should never be crawled and are not already indexed. If filtered URLs are already in the index, let Google crawl them long enough to see a noindex or canonical first.