Treat each filter combination as a product decision first
Faceted navigation can create many URLs from color, size, price, brand, sort order, and availability controls. The useful question is not how to block them all; it is which combinations represent a stable, useful destination that should be crawlable and which are merely temporary interface states. Google’s canonicalization guidance recommends consistent preferred-URL signals, while its robots guidance frames robots.txt as crawl management.
Inventory parameters and paths produced by every filter. For each, decide whether it changes content materially, whether users can bookmark it, whether it has a useful title and page body, and whether it should receive internal links or sitemap inclusion. Keep the output small enough to audit. A filter URL that is neither a valuable landing page nor private may still need a canonical or other handling; do not let a generator decide strategy accidentally.
Work through a catalog example
A furniture shop has /chairs/, then combinations such as /chairs?colour=blue&material=velvet&sort=price. It decides that a curated /chairs/blue/ page is a useful category with unique copy and products, while sort order and multiple ad-hoc combinations are not. Link to and sitemap the curated category. Keep sort variations out of the sitemap and internal navigation; canonical genuinely equivalent sort views to the stable category if appropriate. Do not canonical a “blue velvet chairs” page to /chairs/ if the filtered result is materially different and intended to serve a distinct need.
Test actual generated URLs with filters in different orders, empty results, pagination, and tracking parameters. Inspect the final canonical and status, then see whether a crawler can discover each URL through anchors. Internal links for crawl discovery covers that route evidence. Consider robots rules only after you understand the URL patterns; a broad disallow can block a URL whose canonical or noindex instruction you expect a crawler to read.
Avoid accidental crawl traps
Do not create unlimited calendar, price-range, or query-string variations with ordinary crawlable links. Do not hide a useful category behind JavaScript controls with no stable URL. Do not use robots.txt as access control. Do not populate XML sitemaps with every possible combination. The XML sitemap audit checklist gives a practical inclusion test.
Use the AI search readiness checker to inspect one public response and its raw links. It cannot enumerate all facet combinations or promise a crawl-budget outcome. Re-audit when merchandising, URL formats, or filter controls change.
Revisit the decision as inventory changes
A filter that is low value today may become a curated landing page later, and the reverse can also happen. Keep the selection rules close to merchandising and engineering ownership. Test representative combinations after new attributes, sorting controls, or URL rewrites are released so accidental growth is caught where it begins.
FAQ
Should all filter URLs be blocked?
No. Preserve stable, useful filtered pages; control repetitive or low-value combinations deliberately.
Is a canonical enough for every filter?
No. First determine whether the pages are actually equivalent and whether links create uncontrolled variation.