Faceted Navigation and Index Bloat: Deciding What Deserves a Page
Six colours and five sizes produce 360 URLs per category. A three-tier model for classifying facets, plus the two combination rules that cut bloat by an order of magnitude.
Faceted navigation is the fastest way to turn a 900-product catalogue into 40,000 URLs. Most stores discover this after a crawl shows tens of thousands of pages indexed, none of which rank, all of which consume crawl budget.
The fix is not to remove filters. Filters are useful to customers. The fix is to decide which facet combinations deserve to be pages.
Why the explosion happens
Each filter value multiplies the URL count. Six colours, five sizes, four price bands and three materials produce 360 combinations per category. With 40 categories that is 14,400 URLs, most of which return a handful of products or none.
Left unmanaged, three things follow:
- Crawl budget goes to near-duplicate pages instead of products
- Products end up with multiple internal paths, splitting internal link equity
- Search engines pick a filtered URL as canonical for a category, which is usually not what you want
The three-tier model
Classify every facet, then apply a rule per class.
| Tier | Definition | Treatment |
|---|---|---|
| Indexable | High demand, enough products, stable | Own URL, indexable, in sitemap, linked internally |
| Crawlable | Useful for users, low demand | Real URL, noindex, follow, not linked from crawlable anchors |
| Blocked | Sorting, pagination within facets, nonsense combinations | Robots or canonical to parent |
The middle tier is the one stores forget. A filter combination that gets ten visits a month from real customers is worth having, but not worth indexing.
Choosing the indexable set
Three tests, all of which should pass:
Demand. People search for the combination as a phrase. “Waterproof hiking boots” yes; “waterproof boots size 43 brown” no.
Inventory. The combination returns enough products to be a useful page. Ten is a reasonable floor; three is not.
Stability. The combination will still make sense in a year. Seasonal or promotional facets are not indexable candidates.
Applied honestly, most stores end up with a set of indexable facet pages in the low tens, not the thousands.
Technical implementation rules
Canonical. Every filtered URL should canonical to either itself (if indexable) or the clean category URL (if not). Never canonical a useful filter to an unrelated page.
Robots. Do not block facets with wildcards that also catch legitimate pages. Blocking *?* is the classic mistake — it kills the very pages you decided to index.
Internal links. Indexable facet pages need crawlable links from the category page. A page reachable only through a JavaScript filter UI will not be crawled reliably, however indexable it is.
Sitemap. Indexable facet pages belong in the sitemap; the rest do not. A sitemap containing every combination trains crawlers to waste time.
Page titles. Each indexable facet needs a distinct title that reads like a search query: “Waterproof Hiking Boots” rather than “Boots – Waterproof – Brown”.
Combination rules
The URL count explodes because filters combine. Two practical restrictions:
- Allow at most two facets per URL for indexable pages. Three or more is almost never a phrase people search
- Never allow two facets of the same type — “colour=red&colour=blue” produces one page per permutation
These two rules alone typically cut index bloat by an order of magnitude.
What to do about existing bloat
If the store already has tens of thousands of indexed filtered URLs:
- Export indexed URLs and segment by parameter pattern
- Identify which ones have impressions or clicks — usually a small fraction
- Keep those indexable, with proper titles and internal links
- Set the rest to noindex follow, and let them drop out gradually
- Only use robots.txt for parameters that are purely functional
Do not delete or 410 thousands of URLs at once. A slow reduction is safer, and it lets you observe whether anything valuable was in the set.
The customer-facing side
Every technical rule above has a UX counterpart. A facet system that is fast, keeps state in the URL, and shows product counts per option is better for customers and simpler for SEO — because customers and crawlers follow the same paths.
Which is worth remembering: the reason filter URLs rank is that a filtered result page is often exactly what a shopper wants. The goal is not fewer pages; it is fewer pages that exist only as a side effect of a UI control.
Category-level structure is covered in the category page notes; this piece is the parameter layer underneath it. Both usually get audited together under technical SEO, because the crawl data that shows the problem is the same crawl that shows the fix.