UX / Filtering
Ecommerce Faceted Search: Filters That Convert
// 400 results is a problem; the right six is a sale
By Fracto Solutions
October 7, 2026
The short answer
Faceted search lets shoppers narrow a large result set by attributes like size, color, price, and availability. Good filters are relevant to the category, combinable, show live result counts, and stay applied as the shopper refines. Done well, they turn an overwhelming list into the handful of products someone actually wants.
78%
51%
20%
What is faceted search, and why does it matter?
Faceted search is filtering by product attributes, size, color, brand, price, rating, availability, so a shopper can carve a large result set down to what fits their need. It matters because a wall of 400 products converts far worse than a focused set of six.
Browsing and searching both routinely return more products than anyone will scroll. Facets are how shoppers take control: each filter applied is another statement of intent, which is why this traffic is worth protecting. Constructor, analysing 609 million searches across 113 retailers, found searchers convert at 2.5 times the rate of non-searchers while being only 24% of visitors. A store that filters well feels effortless; one that filters badly pushes those shoppers to a competitor. This is the detailed companion to our site-search pillar.
The industry is not doing this well. Baymard’s product list and filtering benchmark, covering 170+ leading US and European sites against more than 21,000 filtering and product-list parameters and last updated in September 2025, rates 58% of desktop sites and 78% of mobile sites as poor to mediocre on product list UX. Filtering is one of the clearest places where a store can be visibly better than most of its competitors.
Which filters should an ecommerce store offer?
Offer the filters that match how people actually choose in your category, and no more. Universal filters like price, availability, and rating apply everywhere; category-specific ones (size and color for apparel, capacity for appliances) should match the real decision criteria.
Too few filters strands shoppers in a long list; too many or irrelevant ones create paralysis. Most stores err on the thin side: Baymard found 51% of sites do not provide all five essential filter types it tests for, price, user ratings, color, size and brand, and 69% do not offer all four essential sorting options. Check those first, then add only the facets that genuinely drive the decision in your category.
Filter type
Example
When to include
Price
Slider or sensible ranges
Almost always; a near-universal decision factor
Availability
In stock only
Always; never make someone click a sold-out product
Rating
4 stars and up
When you have enough reviews to populate it
Category attribute
Size, color, capacity, material
When it is a real buying criterion for that product
Brand
Filter by maker
Multi-brand catalogs where brand drives choice
Match filters to how buyers actually decide in the category; relevance beats quantity.
How should filters behave?
Filters should be combinable, update results live, show how many products each option will return, and stay visibly applied so shoppers can remove them one at a time. Behavior matters as much as which filters you offer.
Mobile deserves special attention, because cramped or hidden filters are a frequent conversion leak and the benchmark gap between desktop and mobile is wide. The same discipline applies here as in mobile conversion optimization: make the refined, high-intent path effortless on the smaller screen.
Combinable: applying color and size and price together should narrow, not reset. Baymard records 14% of sites failing to let shoppers combine multiple values within the same filter type, so if yours cannot, you are in the minority that does.
Live counts: show how many results each option yields so shoppers avoid dead ends.
Sticky and visible: keep applied filters on screen with easy one-tap removal. 20% of sites give no overview at all of which filters a shopper has applied, which is how people end up confused by a result set they built themselves.
Mobile-friendly: this is the big one. Baymard rates 58% of desktop sites poor to mediocre on product list UX, against 78% on mobile. Filtering is where mobile breaks, so treat it as a first-class flow with a full-screen panel, not a cramped afterthought.
What are the most common faceted-search mistakes?
The usual culprits: filters that reset your other selections, options that lead to zero results, no result counts, and filtering that is painful on mobile. Each one quietly sends a motivated shopper somewhere easier.
Watch especially for filter combinations that dead-end at zero results, the filtering equivalent of a no-results search. If a shopper can select options that cannot coexist, either prevent it or recover gracefully with suggestions, rather than dropping them on a blank page after they did exactly what you invited them to do.
We tune filtering around how your buyers actually choose, then prove it in the funnel.
We look at which filters get used, which combinations dead-end, and where mobile shoppers drop out, then fix the facets and behavior that are costing refined, high-intent sessions. The point is to make the fastest path to the right product effortless.
That is the same measurement-first approach behind our conversion optimization work: we agree the metric before the test runs and report the result at the confidence level the test actually reached, rather than the headline that reads best.
// we would rather show up in your reporting than in your inbox
Frequently asked questions
How many filters should I offer?
Enough to match how buyers decide in your category, and no more. Baymard tests for five essential types, price, user ratings, color, size and brand, and found 51% of sites do not offer all of them, so most stores are short rather than overloaded. Start with those five plus availability, then add only the attributes that genuinely drive choice in your category. Too many irrelevant filters cause paralysis; too few strand shoppers in a long list.
Should filters show result counts?
Yes. Showing how many products each option returns prevents shoppers from selecting combinations that dead-end at zero results, and helps them refine confidently.
Why does filtering break on mobile?
Because filters often get crammed into a tiny control or hidden entirely, and the benchmark data shows how widespread this is: Baymard rates 78% of mobile ecommerce sites poor to mediocre on product list UX, against 58% on desktop. Mobile is usually most of your traffic, so treat mobile filtering as a first-class flow with a full-screen, easy-to-use panel.
What happens when a filter combination has no results?
Ideally prevent impossible combinations, or recover gracefully: tell the shopper, relax the least important filter, and suggest close alternatives, rather than showing a blank page.
Sources
Product list and filtering performance: Baymard Institute product list and filtering UX benchmark, 170+ leading US and European ecommerce sites assessed against more than 21,000 product list and filtering parameters, with 16,000+ best and worst practice examples. 58% of desktop sites and 78% of mobile sites rate poor to mediocre on product list UX. 51% of sites do not provide all five essential filter types (price, user ratings, color, size, brand). 69% do not offer all four essential sorting types (price, user rating, best-selling, newest), on both desktop and mobile. 20% provide no overview of the filters a user has applied. 14% do not let users combine multiple values within the same filter type. Published 22 August 2024, last updated 9 September 2025, so these figures are current as of that update rather than as of today. Baymard: the current state of product list and filtering
Searchers versus non-searchers: Constructor, “Beyond Relevance”. 609 million shopper searches generating $9.8 billion in search-driven revenue across 113 leading ecommerce sites, October to December 2024. Press coverage dates it to 6 March 2025; Constructor publishes no date on the report itself. Searchers were 24% of visitors but drove 44% of revenue and converted at 2.5x the rate of non-searchers. Constructor’s own report page elsewhere says “nearly three times”, so the publisher is not internally consistent; we cite the lower figure. Constructor study, 6 March 2025
Stated as Fracto’s position, not research: matching the filter set to how buyers decide in a given category, and the guidance that filters should stay visibly applied and removable one at a time. These reflect our client work and Baymard’s qualitative guidelines rather than a single quantified study, and we label them as practice.