Faceted Navigation SEO: Which Filters to Index, and Which to Block

The first thing I check on a large ecommerce audit is how many URLs Google is actually crawling versus how many pages the site really has. Nine times out of ten the crawl number is far too high, and faceted navigation is why.

Faceted navigation is the filter panel on a category page: colour, size, brand, price, rating. Useful for shoppers. Quietly brutal for search, because every filter usually writes itself into the URL, and combinations multiply fast.

I have cleaned this up on real sites, and the fix is never "noindex everything". It is a decision, filter by filter. Here is how I make that call, the numbers behind it, and the one mistake that keeps index bloat alive no matter how many tags you add.

What is faceted navigation?

Faceted navigation lets a user narrow a listing by several attributes at once. On a shoe category you might filter by brand, size, colour and price, all on the same page, in any combination.

It is different from your normal menu. Your top navigation is hierarchical: a fixed tree of category and subcategory pages you built on purpose. Faceted navigation is dynamic: it generates a filtered view of the same product set on the fly, and usually a new URL with it.

You see it everywhere large catalogues live. Amazon, Zappos and Best Buy for retail. Booking.com and Expedia for travel, filtered by date, price and stars. Zillow and Rightmove for property, filtered by location, beds and amenities. The pattern is the same: one listing, many filters, an enormous number of possible URLs.

That last part is the whole problem.

Why is faceted navigation an SEO problem?

Filters are not the enemy. Crawlable filter combinations are. The moment each filter state gets its own link and its own URL, one category page stops being one page.

The maths is simple and unforgiving. Take a single category with five filters, each holding ten options. Each filter is either off or set to one of ten values, so eleven states. Five filters combined is 11 to the power of 5.

That is 161,051 URLs from one category page. Add multi-select filters and sort parameters and it runs into the millions. Now picture a few hundred categories doing the same thing.

Four concrete things go wrong when Google is left to crawl all of it:

• Crawl budget waste. Google gives every site a finite crawl allowance. Spend it on ?colour=green&size=9&sort=price-asc and your new products and real category pages get crawled slower.
• Index bloat. Thousands of thin, near-identical filter pages land in the index. A bloated, low-value index drags on how Google sees the whole domain, and that is a site-level quality problem, not a per-page one.
• Duplicate content and cannibalisation. /shoes?colour=black and /shoes?colour=black&sort=newest are nearly the same page. They compete with each other and with /shoes for the same terms.
• Link equity dilution. Internal links pour authority into millions of filter URLs instead of concentrating it on the category pages you actually want to rank.

None of this is theoretical. It is the single most common cause of "we have 20,000 products but Google has indexed 400,000 URLs" that I see on audits.

The four levers, and what each one actually does

Before the decision, you need to know your tools honestly, because they are not interchangeable. There are four, and most guides blur them together.

1. Index and optimise

Let the page get crawled, indexed and ranked. You want this for a small number of filter pages that have their own search demand, a "Nike running shoes" page people genuinely search for. For those, a filter URL is an asset, not a liability.

2. rel=canonical to the parent

Keep the filtered URL, but point its canonical at the clean category page. This tells Google the parent is the real version and stops near-duplicates competing. Google says using canonical this way may, over time, reduce how often it crawls the non-canonical versions. Two honest limits: canonical is a hint Google can ignore, and it does not save crawl budget the way a block does, Google still has to fetch the page to read the tag.

3. robots.txt disallow

Block Googlebot from crawling the URL pattern at all. This is the bluntest and, for crawl budget, the most effective lever. Google's own guidance is direct: use robots.txt to disallow crawling of faceted navigation URLs, because there is often no good reason to allow it, and it consumes server resources. It calls canonical and nofollow "generally less effective in the long term" than blocking. The catch: a disallow cannot remove a URL that is already indexed.

4. noindex, follow

Let the page be crawled, but add a noindex tag so it drops out of the index while still passing link equity through its links. This is the right tool for clearing out filter pages that are already bloating the index.

Here is the summary I keep in my head:

Read the two "Yes" cells carefully. Nothing does both jobs at once. That is exactly where people go wrong.

The mistake that keeps index bloat alive

The most common faceted-navigation fix I have to undo is this one: someone disallows the filter URLs in robots.txt and adds a noindex tag, expecting the URLs to disappear from Google.

They do not. The two cancel each other out.

The reason is mechanical. If a URL is blocked in robots.txt, Googlebot never fetches it, so it never sees the noindex tag inside. And a blocked URL can still get indexed from internal or external links pointing at it, which is why you end up with the classic Search Console warning: "Indexed, though blocked by robots.txt."

Google states the rule plainly: for a noindex rule to be effective, the page must not be blocked by a robots.txt file. If the crawler cannot reach the page, it will never see the rule, and the page can still appear in search.

So the sequence matters:

1. To clear existing bloat, allow the crawl and apply noindex. Wait until the URLs drop out of the index.
2. Only then, if you want to stop wasting crawl budget on that pattern going forward, add the robots.txt disallow.

Block first and you freeze the bloat in place. This is why I say most index bloat is a canonicalisation and crawl-control problem in disguise, not a "just noindex it" problem.

Which filters should you index, block, or canonicalise?

This is the decision the plain guides skip. You do not treat all filters the same, because they do not all have the same value. Run every facet through one test: does this filter combination have its own search demand, and can it stand as a page in its own right?

Here is how that plays out on a typical catalogue:

The judgement calls sit in the top three rows. My rule: if a filtered page pulls real search volume and you can genuinely make it better than the bare category, promote it to a proper, indexable landing page and link to it in your navigation. If not, it is a near-duplicate, so canonicalise it. Everything below that, the sliders, sorts and session junk, is infinite and worthless to search, so block it in robots.txt.

The mistake I see most often is over-blocking, killing off brand or attribute pages that actually had demand. Measure before you swing the axe.

Faceted navigation best practices

Putting it together, here is the checklist I work through on a faceted-navigation cleanup.

1. Measure first. Pull Crawl Stats and the Page indexing report in Search Console, and ideally the server logs. Find out where crawl budget is actually going before changing anything.
2. Find the demand. Check which filter combinations have real search volume. Those become curated, indexable landing pages. Everything else is a candidate for canonical or block.
3. Set clean canonicals. Point near-duplicate filter URLs at their parent category. One clean canonical per cluster.
4. Disallow the junk. Block sort, view, per-page, price-slider and session parameters in robots.txt so Google never spends budget on them.
5. Clear existing bloat correctly. For filter URLs already indexed, apply noindex with the crawl still allowed. Deindex first, then block later if needed.
6. Control the links you emit. For a brand new build, the cleanest option is filters that do not create crawlable URLs at all, using JavaScript or URL fragments. Google does not crawl or index URL fragments, so #colour=red has no crawl impact by design.
7. Keep a consistent parameter order. ?colour=red&size=9 and ?size=9&colour=red are the same page to a human and two URLs to Google. Fix the order so you are not manufacturing duplicates.
8. Monitor. Re-check Crawl Stats and index counts after Google reprocesses. This takes weeks, not days.

None of these is exotic. The skill is applying the right one to the right facet, which is why the decision matrix comes before the checklist.

How do you diagnose it on your own site?

You do not need a crawler to spot it, though one helps. Two places tell you almost everything.

In Search Console, the Crawl Stats report (Settings, then Crawl stats) shows how much of Google's crawling goes to parameter URLs. If a large share of requests are ?-heavy filter URLs, that is your budget leaking.

The Page indexing report is the other tell. Two statuses are the signature of filter bloat: "Crawled - currently not indexed" and "Duplicate without user-selected canonical". A pile of parameter URLs under either one means Google is crawling your filters, deciding they are not worth indexing, and wasting the trip.

A quick site:yourdomain.com search with a filter parameter (site:yourdomain.com inurl:sort=) gives you a rough, free count of how much has already leaked into the index.

If those reports are ugly, treat it as a technical SEO problem to be diagnosed, not a setting to flip. The order of operations, measure, decide per facet, then act, is what separates a cleanup that recovers crawl budget from one that quietly deletes pages you wanted to keep.

Final take

Faceted navigation is not a problem you solve with a single tag. It is a set of decisions: index the few filter pages with real demand, canonicalise the near-duplicates, and block the infinite junk in robots.txt, in that order.

Get the order wrong, block before you deindex, and you lock the bloat in place. Get it right and you hand Google a smaller, cleaner set of pages to crawl, which is exactly what a large site needs to keep ranking.

If your site has more URLs in the index than products on the shelf, faceted navigation is almost certainly why. Start by measuring where the crawl is going, not by noindexing everything in sight.