Services · By journey
A Search Audit Finds the Results Nobody Ever Hears
For the catalog, library, or listing site where finding things is the product, and where filters were built for mouse users first and last.
What We Keep Finding
Search interfaces fail in a signature sequence. The user types, results update live, and the screen reader says nothing, so as far as they can tell the search did nothing. They apply a filter, the count changes silently somewhere above. They land on "no results", announced as emptiness with no suggestion of what to do next.
The announcement half has a detail that decides whether it works at all. It is not the results list that needs announcing, it is the count, and a region that updates only a number will read out "60" on its own unless it is set to read the whole message. "60 of 340 shown" is what a person needs to hear, and getting that is one attribute. There is also a boundary worth knowing before anybody writes a finding. If your filter reloads the page, the announcement rule does not apply, because the page change carries the news. It only bites when results update in place.
Facets pile on operability problems. Checkbox trees with no keyboard path, sliders that only drag, and "show more" toggles that expand into the void. Two corrections there. A drag-only slider fails the keyboard rule at Level A before it fails anything newer, which is the more serious finding and the one to lead with. And a row of applied-filter chips each labelled "Remove" fails the name-and-role rule rather than the link rule, because a chip is a button and buttons do not get the in-context allowance links do.
One thing worth claiming rather than fixing. Site search is one of W3C's published ways of satisfying the Level AA rule about giving people more than one route to a page. So a catalog with working search plus one other route already meets it, and the audit is the thing that establishes whether the search works. Breadcrumbs, for what it is worth, do not count toward that rule, whatever your last report said.
What We Check
- Search and hear the results, so the count is announced on every update
- Check the announcement carries the whole message rather than a bare number
- Operate every facet by keyboard, so checkboxes, ranges, and toggles
- Check applied-filter chips can be reviewed and removed by name, not by "Remove"
- Test the empty state for guidance rather than absence
- Verify sort changes and pagination announce what happened
- Check the search landmark is marked up so a screen reader can jump straight to it
What You Get
Every service on this site runs the same three-pass engine: an automated scan, an expert review of all 55 WCAG 2.2 A and AA rules, and a hands-on session with a professional blind screen-reader user. You get one report with every finding screenshotted, ranked by user impact, and linked to its fix. Your team fixes, we verify: the re-audit is half price within 3 months.
The format is not a mystery either. Read the sample report before you spend anything.
The Honest Limit
One honest limit, and a note on how this gets scoped. Search relevance is your engine's business. We audit the interface around it. Whether the right products come back is yours, and whether anyone can hear them come back is ours. On scope, a search journey is states rather than URLs. Results, filtered results, an empty state and a second page is four of your 10, even though it may all be one address. List them as states when you pick your pages and the count works out honestly.
What It Costs
Rapid Audit: $499, up to 10 pages you pick, report in 5 business days. Standard Audit: $1,499, up to 25 pages in 10 business days. Flat rates, no discovery calls, and a real blind screen-reader user on every engagement. Pick your pages. We bring the humans.
Worth Reading Next
Related on this site
Guides, checklists, tools, and terms that go with this service.