Start with a real category and its available products. A blanket “index every filter” or “block every parameter” rule can hide valuable pages or leave a much larger crawling problem untouched.
Map URL behavior before changing rules
Collect examples of a category, one filter, several filters, sorting, pagination, tracking parameters and an empty result. Repeat the same selection in a different order. Does it create another URL for the same products?
Record the response status, canonical, robots rules, indexability, internal links and whether Google has actually crawled the URL. A crawler's “indexable” label describes technical settings, not Google's decision to index it. Where available, server logs show which patterns consume requests.
Keep this investigation separate from writing a better category page. One concerns URL behavior; the other concerns the usefulness of the destination.
Choose a landing page because the selection is useful
Suppose a bedding store sells linen duvet covers. That material-and-product combination might justify a stable landing page if buyers search for it, there is a meaningful selection and the page can stay useful as stock changes. A sort-by-price URL usually represents the same collection in another order.
Search demand is one input, not an automatic approval. Review the actual results in the target country, the product range, commercial relevance and overlap with an existing category. A filter with two conditions can be valuable; a single filter can be useless. The number of parameters is not a quality score.
Give selected pages a deliberate place in the site: a consistent address, suitable title, useful selection and relevant internal links. Do not create a second landing page when a current category already does the job.
Write down the decision for each URL pattern
This example is a planning worksheet, not a robots.txt configuration to paste into a store. Test the actual platform behavior before applying a rule.
Canonical, noindex and robots.txt do different jobs
Canonical is a signal for duplicate or very similar pages. It is not a command to transfer every filtered page to the parent category, especially when the pages serve different needs.
Noindex needs a crawl. Google must fetch the page to see that instruction. Blocking it in robots.txt at the same time can prevent that. Robots.txt controls crawling and does not reliably remove an already known URL from search.
Google's faceted navigation guidance distinguishes useful filter pages from unwanted crawl paths. Its recommendations include consistent filter ordering and correct handling of invalid combinations. Use those principles with the store's existing URL history; removal from the index and future crawl prevention may require different stages.
Give developers examples and an acceptance check
A useful task says which pattern changes, why, and what must remain unchanged. Include before-and-after URLs, expected HTTP response, canonical, crawl permission and links. Ask the developer to show the rendered result as well as the source response.
After release, repeat the same sample and check that shoppers can still filter, clear selections and reach products on mobile. Follow crawling and indexed-state evidence over time; a reduced URL count alone is not proof of better sales or rankings.
I cover this work within e-commerce SEO analysis, with recommendations matched to the platform and the developer's implementation scope.
Verify your filter URL rules
- Approve indexable filter combinations based on search demand, stable inventory, and a clear difference from the parent category.
- Compare parent, filtered, and reordered-parameter URLs for status, noindex, declared canonical, and Google's selected canonical in indexed URL Inspection data.
- Crawl links and pagination after deployment; use the live test to check current accessibility, not to confirm Google's canonical choice.
Faceted navigation SEO FAQs
Which filter pages should we let Google index?
Keep combinations that match distinct searches and support useful, stable landing pages. A material-and-size combination may deserve its own page; every sort order and price slider setting usually does not. Compare each candidate with its parent category to determine whether it serves a meaningfully different shopping need.
Should we use robots.txt, noindex, or canonical for filter URLs?
Use canonical hints for genuinely duplicate URLs and a crawlable noindex for pages that should not appear in search. Robots.txt controls crawling and can prevent Google from reading either signal. For already indexed URLs, let Google process the removal signal before blocking crawling. Plan index removal and crawl reduction separately.
Can we fix filter SEO without making shopping harder?
Usually, yes. Keep the shopping controls, but separate them from the URLs promoted as search landing pages. The SEO handoff should define indexing, canonical, and linking rules by URL pattern. A developer implements parameter handling and link generation; the SEO reviewer checks representative combinations after release.