The problem filters create
Faceted navigation lets shoppers narrow a category by brand, size, color, price and more. With a handful of facets, the number of possible combinations explodes: four filters with ten values each can produce thousands of unique URLs. Most show the same products reordered or a thin subset. Search engines may waste crawl effort on them, index low-quality pages, split ranking signals across duplicates and miss your important products.
The opposite problem
Blocking every facet is also a mistake. Some filtered pages match real searches, such as “black leather office chairs” or “Nike running shoes size 10.” These deserve to be proper, indexable landing pages with unique titles and content.
The goal is selectivity, not blanket blocking.
Step 1: decide which facets deserve pages
Use search data and business logic.
| Facet type | Indexable? | Reasoning |
|---|---|---|
| Brand | Often yes | People search by brand within categories |
| Product type or style | Often yes | Matches common queries |
| Color or material | Sometimes | Only where demand exists |
| Size | Usually no | Highly specific and low search volume, with exceptions |
| Price ranges | Usually no | Arbitrary ranges rarely match queries |
| Sort order | No | Same products, different order |
| Availability | No | Changes constantly |
| Multi-facet combinations | Rarely | Only if proven demand |
Check keyword research, internal search data and past conversions to find the exceptions.
Step 2: choose the right controls
No single tool fits everything. Combine them thoughtfully.
- Clean URLs for indexable facets. Give valuable combinations static-style URLs, a unique title, heading and some helpful content, and include them in the sitemap.
- Canonical tags. Point low-value filtered URLs to the parent category. Canonicals are a hint, not a command, and crawlers still spend effort visiting those pages.
- Noindex. Use for pages that should be crawled to discover products but not indexed. They still consume crawl budget.
- Robots.txt disallow. Prevents crawling, which saves crawl budget, but blocked pages cannot pass signals, and search engines cannot see canonical or noindex tags on pages they cannot fetch. Use for parameters with no value, like session IDs and sort options.
- Non-crawlable filters. Implement filter controls with JavaScript that updates results without exposing crawlable links, or use form elements, while keeping indexable facets as regular links.
- Parameter order and consistency. Make sure the same filter combination always produces the same URL, regardless of the order chosen.
Step 3: keep internal links clean
Internal links should point to preferred URLs. Avoid linking to sort or deep filter combinations from navigation, breadcrumbs or related-product blocks. Check that your canonical, sitemap and internal links agree.
Step 4: handle empty and thin results
Filter combinations that return zero or very few products should not be indexable and should return a helpful message, not a soft error. Do not generate pages for combinations without inventory.
Step 5: handle pagination within facets
Each paginated page of an indexable facet should have a self-referencing canonical and crawlable links, as with categories. Avoid infinite scroll without link-based fallbacks, so search engines can reach all products.
JavaScript and performance
If filters load results with JavaScript, confirm search engines can render them and that indexable facets have real URLs. Check speed, because filter interactions that reload large pages hurt the experience. Fast, clear filtering improves conversion as well as SEO.
Audit your current state
- Crawl the site and count unique URLs by parameter.
- Review server logs to see where search engine bots spend time.
- Check indexed pages in Search Console for parameter URLs.
- Identify the facet pages that already get impressions.
- Compare against your intended rules, and fix the gaps.
Common mistakes
- Using only canonical tags and assuming the crawl problem is solved.
- Blocking in robots.txt while expecting noindex to work.
- Letting multi-select filters generate endless combinations.
- Leaving internal sort links crawlable.
- Index bloat from empty combination pages.
FAQ
Should I noindex all filtered pages?
No. Some filter combinations match real searches and deserve indexable pages. Noindex or otherwise control low-value ones.
Is a canonical tag enough?
Not usually. Canonicals consolidate signals but do not stop crawling. Combine them with crawl controls and clean internal linking.
How do I know which facets people search for?
Use keyword research, Search Console queries, internal site search data and conversion data to find high-demand combinations.
Can robots.txt and noindex be used together?
If robots.txt blocks a page, crawlers cannot see its noindex tag. Choose the control that matches your goal for each URL pattern.
TURN INSIGHT INTO ACTION
Let’s talk about your next step.
Share your workflow, your current systems and what you want to improve. We’ll help shape a practical way forward.
Discuss your project ↗