What is faceted navigation in an online store?

You know it even if you do not know its name: it is the column of filters on a category, the one that lets you narrow down to running shoes in size 9, black, from a particular brand. Each attribute you can filter by is a facet, hence the name. For the shopper it is the fastest way to get to the product. For a crawler it is something else: every filter combination can be a new URL, and a single category can manufacture thousands. It matters because those URLs compete for the time search engines spend on your store, and because almost all of them show the same products as the category they came from.

How it works

Google defines it as a common feature of websites that lets visitors change how the items on a page are displayed, whether products, articles or events, and illustrates it with a parameter URL such as items.shtm?products=fish&color=radioactive_green&size=tiny.⁠^{[1]Google, crawling documentation. Improve crawling of faceted navigation URLs} The problem it describes is one of crawling: because those URLs look new and the crawler cannot know whether they are useful without requesting them, it often visits a very large number of filter URLs before concluding they are not useful, and the time spent on them is not spent on new, useful URLs.⁠^{[1]Google, crawling documentation. Improve crawling of faceted navigation URLs} The detail of that split is the crawl budget.

What is rarely calculated is how fast the combinations grow. A fictional example calculation: a category with a size filter of 9 values, a color filter of 6 and a brand filter of 4.

The second row is the one that tends to surprise. As soon as a filter accepts several values at once, each facet stops having "9 options" and has all of its possible subsets. No store has 524,287 distinct listings to show: most of those combinations repeat products or return none.

Google separates two decisions.⁠^{[1]Google, crawling documentation. Improve crawling of faceted navigation URLs}

• For filter URLs you do not need in the search engine, it suggests blocking their crawling with robots.txt, with disallow rules for each filter parameter, or moving the filters into URL fragments (#color=green), because Google Search generally does not support fragments for crawling or indexing. Pointing filtered URLs at the category with a canonical can reduce crawling over time, and putting rel="nofollow" on links to filtered results can help, but Google considers these two routes generally less effective in the long run.
• For the ones you do need crawled, it asks you to use the standard & separator between parameters, keep a logical, fixed order of filters in the URL, not repeat a filter or allow nonsensical combinations, and return a 404 when a combination has no results.

The question for each facet is therefore whether any of its combinations answers a search someone really makes, and so deserves a category page of its own with a stable URL, or whether it only serves to narrow the list someone is already looking at.

In a store

In Shopify, storefront filtering is the recommended method for filtering products in a theme, and it works on collections as well as on search results. Applied filters travel in the URL as parameters of the form filter.scope.attribute=value, where the scope is p for product filters and v for variant filters: for example, filter.p.product_type=shoes or filter.v.option.color=red. You can filter by availability, price, product type, vendor, tags, category, variant options and metafields. Different filters combine with "and"; several values of the same filter, with "or", and you can create up to 25 filters.⁠^{[2]Shopify.dev. Storefront filtering}

Here is what the URL map of a single collection would look like in a Shopify store, in a fictional example:

Shopify already ships two defenses in its default robots.txt. The Disallow: /collections/*+* rule stops crawling of filtered collection pages, because they duplicate the content already on the collection page, and the Disallow: /collections/*sort_by* rule stops crawling of sorted collections, because they return the same products in another order.⁠^{[3]Shopify Help Center. Editing robots.txt.liquid} What your store does with the filter. parameters is worth checking in your own file.

How to check it

1. Apply two filters in one order and then in the opposite order. If the URL changes, each combination has at least two addresses; Google asks for a fixed order.⁠^{[1]Google, crawling documentation. Improve crawling of faceted navigation URLs}
2. Combine filters until no product is left and look at that page's status code (in the browser's developer tools, network tab). Google asks for a 404 in that case.⁠^{[1]Google, crawling documentation. Improve crawling of faceted navigation URLs}
3. Open your /robots.txt and look for which rules cover the filter, sort and pagination parameters. In Shopify, check whether the default rules for filtered and sorted collections are still there.⁠^{[3]Shopify Help Center. Editing robots.txt.liquid}
4. Review Search Console's page indexing report. Filter URLs tend to show up under statuses such as "Blocked by robots.txt" or "Duplicate without user-selected canonical"; the goal Google sets is for the canonical version of every important page to be indexed.⁠^{[5]Search Console Help. Page indexing report}
5. Decide facet by facet. Note which combinations you want findable, with their stable URL, and block or pull the rest out of crawling with one coherent rule, not three methods at once.

Related concept: duplicate content

Faceted navigation is one of the most common causes of duplicate content in a store: Google cites the results of sorting and filtering a category among the reasons a site has duplicate content.⁠^{[6]Google Search Central. What is URL canonicalization} The difference lies in which problem weighs more. Duplicate content is mostly a question of which URL counts, and it gets sorted with signals such as the canonical; faceted navigation is also a crawling question, and there the canonical arrives late: Google places it among the less effective routes in the long run.⁠^{[1]Google, crawling documentation. Improve crawling of faceted navigation URLs} The PLP entry explains how all this fits into the product listing.

To keep going, the crawling and indexing reading path orders the rest of the terms.