Faceted Search: The Filters That Help and the Ones That Bury Products
Filters should follow how shoppers think, not how your catalog is organized.
Facets are the filters on the left of a search results page: size, color, price, brand. Done well, they help a shopper narrow a hundred results to the ten they want. Done badly, they bury products under filters that do not match how anyone shops. The difference is whether the facets follow the shopper's thinking or the catalog's structure.
Facets that follow the shopper
The best facets are the ones a shopper would ask for. Size, color, and price are the obvious three, because they are how people narrow a choice in a physical store. A shopper looking at a hundred jackets wants to see the ones in their size, in black, under $200. Three clicks, and the list is ten items long. That is a facet doing its job.
The test for a facet is whether a shopper would use it without being told. If the attribute is something the shopper already thinks about before they search, it belongs in the filter list. If it is something only the merchandising team thinks about, it does not, and putting it in the list adds a click the shopper did not ask for.
The facets that bury products
Facets go wrong when they follow the catalog instead of the shopper. A filter for material with 40 values, a filter for season that splits products by a code nobody outside the buying team understands, a filter for a technical attribute the shopper has never heard of. Each one is a wall between the shopper and the product, and the shopper will not climb it.
The worst case is a facet that hides the product entirely. A shopper who filters by size 10 and gets no results because the product is listed as 10.5 in the catalog has not learned that the product is out of stock. They have learned that your search does not work. The facet did not narrow the choice. It removed it.
Counts, and hiding the empty values
Every facet value should show how many products it matches, and values with zero matches should be hidden, not shown as a dead end. A shopper who sees that size 10 has 12 products and size 12 has 3 can make an informed choice. A shopper who clicks size 12 and gets an empty page has wasted a click and lost trust in the filter.
This is a small detail that matters a lot. Stores that show counts and hide empty values have search pages that feel fast and honest. Stores that show every value, including the ones with nothing behind them, have search pages that feel like a maze. The difference is a few lines of code and a willingness to admit that some values are empty.
How many facets is too many
Five to seven facets is the practical limit for most stores. Beyond that, the filter panel becomes a second catalog, and shoppers stop using it. The facets that survive are the ones with the highest usage in the analytics: size, color, price, and one or two attributes specific to the category. The rest can go into an advanced filter or be dropped entirely.
The way to decide is the data. Look at which facets shoppers actually click, and in what order, and cut the ones nobody touches. A facet that no one has clicked in a month is not a feature. It is clutter, and clutter costs conversions in a way that is hard to measure but easy to feel.
Keep reading
Related guides on this site: query relaxation and what the engine should do when the strict search returns nothing part numbers and SKUs and search for the shopper who already knows what they want manual overrides and the merchandising log nobody reads More from Product recommendations shoppers actually want.
Frequently asked questions
What are faceted search filters?
The filters on a search results page, like size, color, price, and brand, that let a shopper narrow a large set of results to the few they want.
When do facets help?
When the filters match how shoppers think about the product: size, color, price, and the one or two attributes that actually decide the purchase. Too many facets, or facets nobody uses, slow the shopper down.
When do facets bury products?
When a facet splits products in a way shoppers do not understand, or when it hides the products they want behind a filter they would never click. A facet with 40 values is not a filter, it is a wall.
Should facets show result counts?
Yes. Showing how many products match each value, and hiding values with zero results, keeps the shopper from clicking into an empty page. A facet that leads nowhere is worse than no facet.