Faceted navigation: when your filters cost you sales
Faceted navigation is the filtering system of a catalog. Built on URL parameters, it generates a practically infinite space of URLs. Google documents two harms that follow: its crawlers spend heavily on useless URLs, and they take that much longer to discover your new pages.
- Four filters offering ten values each produce ten thousand URL combinations for a single category. The full catalog then runs into the hundreds of thousands.
- Google names two harms: the over-crawling of worthless URLs, and the slowdown in discovering new pages.
- Its main recommendation is to block the crawling of these URLs in the robots.txt file, while keeping one unfiltered listing page.
- Filters passed through a URL fragment have no effect on crawling, because Google generally does not support fragments.
- The commercial cost is not theoretical: a product added to the catalog takes longer to rank, and therefore longer to sell.
On this page
Faceted navigation
Faceted navigation is the filtering system that lets a person narrow the display of a list of products, articles or events by criteria such as brand, color, size or price. Its most common implementation encodes each filter in a URL parameter, which multiplies the possible combinations and creates a practically infinite space of URLs. Google documents this implementation as harmful to the crawling of a site, without the feature itself being at fault.
The two documented harms
This topic looks technical and it is not. It is about the delay between the moment you pay for a product and the moment it starts to sell, which makes it a question of working capital as much as of search ranking.
Google explains the mechanism plainly. Because the URLs created by filters look new, its crawlers cannot tell whether they are useful before crawling them. So they visit a very large number of them before concluding that they serve no purpose. That is the first harm: over-crawling.
The second follows directly. Time spent on worthless URLs is time not spent on your new pages. Google puts it this way: the discovery of new and useful URLs slows down. For an online store, that means a product just put in stock takes longer to appear in the results, and therefore longer to clear inventory you have already paid for.
Google also notes that crawling these URLs consumes significant server resources, because of the number of URLs and the operations needed to produce each page. The hosting cost of a poorly configured catalog is therefore real and recurring, on top of the search ranking cost.
What to settle before committing to a technical project
This project produces nothing visible: it removes a waste. That is what makes it hard to fund and easy to postpone, year after year, while the cost piles up in the time to market.
- How much time passes today between adding a product and its first online sale?
- How many distinct URLs can our catalog theoretically generate, filters included?
- What share of our pages is marked "discovered, currently not indexed" in Search Console?
- How much does our hosting cost, and what share of it serves crawlers on worthless pages?
- If time to market has not dropped in six months, what do we cut?
The answer that holds up starts by measuring the number of URLs generated and the current time to market. A hollow answer proposes to canonicalize everything, promises a ranking gain, or treats the subject as a setting without measuring what it costs today.
Deciding whether your catalog is losing sales because of its URL structure comes down to figures you can pull in a day. A 90-minute consultation settles it, with a written summary your technical team can execute.
The three ways to fix it, in order of effectiveness
Google itself ranks these methods, and the order matters: the last two are described as less effective in the long run. Choosing the wrong one costs a second project two years later.
Block crawling in the robots.txt file
This is the main recommendation. Google writes that there is often no good reason to allow the crawling of filtered listings, since it consumes server resources for no or negligible benefit. The setup consists of disallowing the filter parameters while keeping product pages and one full unfiltered listing page accessible. The implementation cost runs to hours for most catalogs, against months of wasted crawl budget.
Run the filters through a URL fragment
Google generally does not support fragments in crawling and indexing. Filtering built on a fragment rather than a parameter therefore has no effect, positive or negative, on crawling. It is an architecture decision made at the time of a redesign: after the fact, it costs far more.
Canonical and nofollow, as a last resort
Pointing the canonical of filtered pages to the unfiltered version can, over time, reduce the crawl volume of the non-canonical versions. Marking the links to filtered pages as nofollow can help too, but every link to a given URL must carry the attribute for it to work. Google describes both methods as generally less effective in the long run than the previous ones. They are fixes, not solutions, and the cost reappears as soon as the catalog grows.
What stays in-house: the list of filters that actually bring in sales, and those that only serve browsing comfort. It is a commercial decision, not a technical one, and no one outside can make it. What can be delegated: counting the URLs generated, configuring the robots.txt file, checking the canonicals, tracking indexing and recording time to market before and after. A team that names five commercial filters gets a clean setup. A team that asks to keep everything gets the original problem, documented.
If you want to keep some filtered pages
Some combinations deserve to be indexed, because they match a real search demand and therefore identifiable sales. It is a commercial decision before it is a technical one: every combination kept carries a recurring crawl cost. Google then sets precise, demanding rules.
| Rule | What it prevents |
|---|---|
| Use the ampersand as the parameter separator | The comma, semicolon and brackets are hard to recognize as separators |
| Keep filter order always identical | Different URLs pointing to identical content |
| Disallow duplicate filters | The proliferation of equivalent URLs |
| Return a 404 code when a combination yields nothing | Empty pages crawled and indexed endlessly |
Google notes that letting these URLs be crawled increases resource usage on your server and can slow the discovery of your new pages. Keeping filtered pages indexable is therefore a choice whose cost is real and recurring, and it is justified only for the combinations whose search demand and expected revenue you can name. Five chosen combinations are worth more than two hundred left open.
The general framework of crawling is covered in publishing 100 pages at once, mastering a large catalog in controlling a catalog before optimizing pages, technical checks in the technical checks to run on a site, and the role of internal links in how authority moves between your pages.
The general framework for search is set out in a complete guide for SMBs and large companies.
Turning a large catalog into an asset that ranks is at the heart of the Attract customers with SEO and AI goal.
Already running a marketing team? See how we plug in as reinforcement on SEO for large catalogs.
Frequently asked questions about faceted navigation
Why do filters hurt search rankings?
Because they create a practically infinite space of URLs. Google explains that its crawlers cannot know whether these URLs are useful before crawling them, so they visit a very large number of them for nothing. The time spent this way slows the discovery of your new pages, and therefore the delay before a product starts to sell.
What is the method Google recommends?
Block the crawling of filter URLs in the robots.txt file, while keeping product pages and one full unfiltered listing page accessible. Google writes that there is often no good reason to allow the crawling of filtered listings, the benefit being nil or negligible.
Are canonical tags enough?
No, they help but remain a fix. Google describes the canonical and the nofollow attribute as generally less effective in the long run than blocking through robots.txt or using URL fragments. The canonical can reduce the crawl volume over time, without removing it.
What should you do with filter combinations that return nothing?
Return a 404 code. Google explicitly recommends that a combination yielding no product return a page-not-found error, for people as well as for crawlers. An empty page that responds normally keeps being crawled and can be indexed, which consumes crawl budget on content that will never sell.
- Google, Managing crawling of faceted navigation URLs, official crawling documentation, accessed July 2026.
- Google, Optimize your crawl budget, official documentation, updated December 2025.
- Google Search Central, Consolidate duplicate URLs, on the use of the canonical tag, accessed July 2026.
Geneviève puts the strategy for your engagement into action. She leads all our web development projects: Shopify, WordPress and the new ways of building a site with AI. She manages our team of developers and translates your business needs into technical language. She runs your organic search (SEO), your visibility in AI answers (GEO) and your site's conversion rate optimization (CRO). Her work is at the heart of three goals: Attract customers with SEO and AI, Improve your site's conversion, and Strengthen your visibility in AI answers. With Gabriel, she also builds the landing pages for your advertising campaigns. She writes mainly about SEO, AI visibility and web design.
About Falia →
SEO
Paid advertising
Conversion
AI visibility
Web design