Site Search and Filtering: How Customers Actually Find Products
How on-site search and category filtering shape how fast a customer finds a product, the most common filter design mistakes, and a practical checklist.
rabbitclip teamPublished: 5 min read
Short answer
On an e-commerce site, site search is what a customer uses when they already know roughly what they want; filtering is what narrows down a category while they're still browsing. The two serve different behaviour and need different design: search should tolerate a typo and recognise a synonym, while a filter should keep the current selection visible at all times and group a long option list by priority rather than dumping it alphabetically.
Baymard Institute's large-scale product list usability research shows that even major e-commerce sites frequently lose customers midway through search and filtering; the cause is usually a design decision, not a lack of technology.
This piece covers why search and filtering need separate treatment, the most common mistakes, and a practical checklist for a category page.
Why site search should behave differently from a category menu
Site search is a channel where the customer describes a product in their own words, so it needs to tolerate a typo (returning results for 'shoez' as well as 'shoes') and recognise a synonym ('jumper' surfacing a product tagged 'sweater'). A category menu shows a predefined structure and doesn't adapt to the customer's own wording at all.
A stationery brand's search engine had no typo tolerance, so a customer searching 'noteboook' got a zero-result page; that small technical gap made a product that was actually in stock look like it didn't exist.
What gets typed into the search box is also a valuable data source; regularly reviewing the most-searched terms that return nothing surfaces either a product that needs adding to the catalogue or a missing synonym match. Reviewing that list monthly keeps search from becoming a feature that was set up once and then forgotten.
What filtering actually solves
Filtering means narrowing a broad product list within a category down by a criterion the customer sets, price range, colour, size, material. In a furniture category with hundreds of products, the right filter narrows a customer down to five products in thirty seconds; a missing or broken filter leaves them browsing page after page, and they usually give up.
Filter options should be generated automatically from the actual catalogue data; a manually maintained filter list falls behind the moment a new colour or material gets added, and a customer can never find that product through the filter at all. It looks like a small technical detail, but it's the most common reason a filter quietly stops working as a growing catalogue expands.
The most common mistake in filter design
Baymard's filter UI research points to two recurring problems: an applied filter not being highlighted anywhere visible at the top of the page, so the customer forgets which one they set, and a filter group, colour or brand especially, dumped into one long unprioritised list.
A home textiles brand's category page listed fifty fabric types alphabetically; the three or four popular options a customer actually wanted got lost in the middle of the list. Pulling popular options to the top and tucking the rest behind a 'show more' link made the same filter group far more usable.
Applying several filters at once holds its own trap: if a customer selects a colour, a size and a price range together and the result drops to zero, the system should never just show a silent empty page, it should suggest which filter to drop to get results back. That small detail saves the customer from removing filters one by one to find the culprit.
A practical filter checklist for a category page
This checklist applies directly when reviewing the filter side of an existing category page.
- Is the applied filter clearly visible at the top of the page or in the side panel
- Does the remaining product count update instantly once a filter is applied
- Are the most-used filter options (price, size, colour) at the top of the list
- Is a long filter group trimmed behind a 'show more' link
- Can filters be reached one-handed on mobile without scrolling
- Can all filters be cleared with a single tap
What should happen on a zero-result search
A zero-result search page can be the exact moment a customer leaves the site entirely; showing near-matching products, bestsellers or a category suggestion, rather than just 'no results found', keeps them on the site. A typo-tolerant search already prevents most of these cases, but a fallback is still needed for the small share that genuinely return nothing.
A pet products brand's data showed that once a 'bestsellers' block was added to the zero-result page, the share of visitors leaving the site straight from that page dropped noticeably; even a customer who didn't find what they searched for stayed on the site, because there was something concrete to click next.
How search and filtering should differ on mobile
With limited screen space, filters on mobile usually open as a bottom sheet rather than a full page, keeping the product list visible in the background while the customer applies a filter. The search box, meanwhile, should stay fixed at the top and not disappear when scrolling through a category; a customer shouldn't have to scroll back to the top every time they want to search again.
Search and filtering sit closer to the sale than almost any other technical decision on an e-commerce site; a small design flaw can make a product that's actually in stock look invisible. In a discovery call with rabbitclip, we review the search and filter side of existing category pages against this checklist.
FAQ
Is site search the same as a category menu?
No; search should adapt to the customer's own wording, while a category menu presents a fixed, predefined structure.
When does a filter list count as too long?
When a customer has to scroll to find the popular option they want; at that point, popular options should move to the top and the rest should be tucked away.
What should a zero-result search show?
Not just 'no results', but near-matches or bestsellers; that keeps the customer on the site instead of losing them entirely.
How should filters open on mobile?
As a bottom sheet, not a full page; the customer should still sense the product list sitting behind it while applying a filter.
