Your online shop lets customers narrow a range by size, colour, material or price. That helps people buy. It can also leave your website with many more URLs than useful search landing pages.
The right approach is to decide which selections deserve a permanent place in search, then control the remaining browsing combinations. Keep the filters customers need while giving Google a clear route to your important categories and products.
This guide explains how to make that decision, what to ask your developer to change and how to check the result.
Why product filters create an SEO problem
A filter URL might look like /jackets/?colour=black&size=medium&sort=price. Each selection can create another address.
Google’s current guidance on faceted navigation explains that these URL spaces can cause excessive crawling and slow the discovery of useful new pages. Its recommended starting decision is whether those filtered URLs need to be considered for indexing at all.
A UK clothing retailer might want a strong page for waterproof jackets. It probably does not need a separate search landing page for every size, price band and sorting order within that range. That is an illustrative judgement, not a rule about what any particular shop’s customers search for.
Separate search landing pages from browsing options
Start with one important product category. Ask the person responsible for sales or merchandising to review the filters with your SEO specialist and developer.
Our practical recommendation is to consider a selection for a dedicated landing page when it passes these checks:
- Clear buying intent: it answers a recognisable customer need that your broader category does not answer as well.
- A useful range: the selection offers meaningful choice and is likely to remain worth visiting.
- A distinct explanation: you can help customers understand suitability, differences or limitations without repeating the parent page.
- Commercial relevance: the business wants to sell that range and can maintain it.
- Supporting evidence: customer enquiries, your own site-search information or Search Console queries justify investigating it.
Missing query data does not prove there is no demand. Treat it as a reason to investigate, rather than inventing a search-volume estimate.
For the jacket example, waterproof jackets and insulated jackets could warrant separate pages if the ranges and customer needs support them. “Waterproof, medium, black, cheapest first” is more likely to be a temporary shopping selection.
Do not create every combination first and hope the useful ones emerge. Agree a manageable list of pages with a named owner for keeping each range accurate.
Give valuable selections a stable home
Where a selection earns a landing page, choose one dependable URL. That might be an existing subcategory or an approved filtered address. A new folder structure is not automatically necessary.
Google’s ecommerce URL guidance recommends persistent URLs, avoiding temporary session or time-based parameters in internal links, and using the same preferred address in internal links, sitemaps and canonical annotations. It also recommends a self-referencing canonical on indexable pages.
Ask your developer to confirm that the approved landing page opens directly, shows the intended range and is eligible for indexing. Give it a useful title and heading, and enough relevant information to help someone choose. Adding a paragraph of generic keywords above an unchanged product grid does not establish a convincing reason for a separate page.
Link to the selection from an appropriate category or other useful navigation. Google’s site-structure documentation recommends linked routes from categories through to products and notes that Googlebot generally does not submit searches into a website’s search box.
The practical check is simple: can a customer reach the range and its products by browsing, without knowing exactly what to type into your search tool?
Choose the right treatment for the other URLs
Crawling means fetching a page. Indexing means considering and storing it for search. Canonicalisation means choosing the representative URL among duplicate or very similar pages. These are different decisions, so the controls are not interchangeable.
Use canonical URLs for genuinely similar pages
A canonical annotation identifies your preferred representative page. Google treats it as a strong signal, rather than an instruction it must follow. Its canonical guidance applies to duplicate or very similar pages and warns that a robots.txt block does not prevent a URL itself from being indexed.
A different sorting order may still show essentially the same range. A genuinely different subcategory may deserve its own page. Ask your SEO specialist to compare the content before pointing all filtered selections at the parent category.
If you want a selected range to appear independently, check that your platform is not simultaneously declaring the broad parent page as its preferred canonical.
Use noindex when the aim is exclusion from search
A page may remain useful to shoppers without needing to appear in search results. Google supports noindex in a page’s metadata or HTTP response header, but it must crawl the page to see that rule. Google’s noindex documentation explicitly warns against blocking the same page in robots.txt and expecting its noindex rule to work.
For already indexed filter URLs, agree the removal sequence with your developer. If noindex is the chosen method, keep those URLs accessible while Google processes it and verify the outcome before considering further crawling restrictions.
Use targeted crawling restrictions for unnecessary combinations
For filter URLs that do not need search visibility, Google’s faceted-navigation guidance supports targeted robots.txt restrictions. It also describes fragment-based filters as an alternative when separate indexing is unnecessary.
These are implementation choices. Do not copy a blanket rule that blocks every URL containing a question mark: useful pagination, product variants or other important routes may use parameters too. Have the developer test the proposed patterns against both unwanted combinations and URLs that must remain accessible.
If the site already has a substantial crawling problem, address how the filter system generates and exposes URLs as well as its search settings. A canonical annotation alone does not stop a crawler fetching the page.
Protect product discovery and handle empty results
Pagination deserves its own check. Page two of a category can contain products missing from page one. Google’s pagination guidance recommends distinct URLs, links between successive pages and a canonical for each page, rather than declaring page one canonical for the whole sequence.
Before restricting filter URLs, confirm that products remain reachable through the main category route, including later pages. Test any “load more” behaviour as well as the first screen.
For filter combinations that return no results, Google’s faceted-navigation guidance recommends an HTTP 404 at that URL. It extends that advice to nonsensical or duplicate filter combinations and nonexistent pagination pages. A helpful customer message should accompany the response; do not redirect every failed combination to a shared error page.
Keep this separate from decisions about an established category whose stock has temporarily changed. Your team should review what the category still offers customers before applying one automated empty-page rule across the shop.
Ask for evidence before approving the change
A developer handover should show a small, representative set of URLs with the expected behaviour recorded. Include:
- The main category and one approved search landing page.
- A sorting-only URL and an unwanted multi-filter combination.
- An empty combination and a later pagination page.
- A product URL and, where relevant, a product variant used by your shopping feed.
For each, record the HTTP response, crawling permission, any noindex rule and the canonical destination. Also test the customer experience: selecting and clearing filters, sharing the URL, returning with the back button and opening a product.
Use Search Console’s URL Inspection tool to compare live accessibility with Google’s indexed information. The Google-selected canonical is available in indexed data; a live test cannot predict it. A successful live test also does not guarantee indexing.
After release, monitor the approved landing pages and samples of the unwanted URLs. Where your developer has access to server logs, ask whether crawling of unnecessary combinations is reducing. Judge commercial progress through relevant visits and purchases, rather than treating a lower excluded-page count as the objective.
Allow time for processing. Google’s current canonical troubleshooting guidance says pages can remain in a duplicate cluster for up to two weeks after content fixes. That is guidance about re-evaluation, not a promise of indexing or improved rankings within a fortnight.
Start with one category and a clear decision
Choose a category where the filters matter to customers. Identify the selections worth maintaining as search landing pages, agree the treatment of the remaining URLs and test the complete shopping route before rolling the approach out.
BuzzBoost’s SEO services and web development work can bring the search decision and implementation together. If your shop is generating more filtered URLs than useful landing pages, talk to BuzzBoost about reviewing the category structure and filter behaviour.
Featured image: AI-generated editorial artwork, not a photograph of a real BuzzBoost office, client or result.
