A clean scan isn't compliance: what I found on 50 Central Arkansas public websites
Twelve of the public websites I tested this week came back perfect. Zero automated accessibility errors on the homepage. If you run a city, county, school district or state agency website, that's the screenshot you'd want to forward to your boss.
Then I put the mouse away and pressed Tab.
On one of those "perfect" sites, the main navigation shows no sign of where you are. Keep pressing Tab and focus drops into drop-down menus that never open, landing on links nobody can see. On three others, the "skip to main content" link that keyboard users depend on stays invisible when it gets focus. On one of them, it sits 160,000 pixels above the top of the page.
No scanner flagged any of it.
That gap, between what a tool reports and what a person can actually use, is the most important thing for public entities to understand before the federal deadline.
The deadline, briefly
In April 2024 the Department of Justice set WCAG 2.1 Level AA as the technical standard for state and local government websites and apps under Title II of the Americans with Disabilities Act. This April, DOJ pushed the compliance dates back a year with an interim final rule. The standard didn't change.
- State agencies, and cities and counties with 50,000 or more residents: April 26, 2027
- Smaller local governments and special districts: April 26, 2028
Public colleges and universities that are part of state government generally fall under the state date. Health care organizations that receive HHS funding have a parallel Section 504 deadline: May 11, 2027 for those with 15 or more employees, and May 10, 2028 for those with fewer than 15.
April 2027 is less than six months away. Even the 2028 date sounds like plenty of time until you remember most public websites are built on templates, maintained by small teams, and managed through contracts that renew once a year.
What I looked at
I tested the homepages of 50 public-sector websites in Central Arkansas: cities, counties, school districts, state agencies and boards, public colleges, the transit system, water utility, airport, port, library system and tourism bureaus. Every homepage got an automated WCAG 2.1 check. I then walked 16 of them with a keyboard, including all 12 that passed the automated check, and confirmed every keyboard finding with a screenshot before counting it.
Homepages are the most-polished page on most sites, so these numbers are a floor, not a ceiling. I'm not naming sites. The point isn't who's behind; nearly everyone is, somewhere.
What I found
38 of 50 homepages (76 percent) had at least one automated failure. Fourteen had 10 or more. The worst had 138 on the homepage alone.
Two problems made up about half of everything:
- Links and buttons with no name (22 sites). Mostly icons: a Facebook logo, a search magnifier, a menu button with no text behind it. On a screen reader, a row of social icons reads as "link, link, link, link."
- Text that doesn't contrast enough with its background (21 sites). Usually a brand accent color: a light teal link, a pale label, a gray heading on white.
Both are template problems. They sit in the header, footer or design system, so they fail on every page, and fixing the template once fixes them everywhere.
A third of the "clean" sites still failed. Of the 12 homepages with zero automated errors, at least four fail WCAG 2.1 AA on keyboard problems I confirmed by screenshot: invisible skip links, missing focus indicators, and focus that wanders into hidden menus. Three passed every keyboard check I ran. The other five I couldn't confirm either way.
A quarter of the sites run an accessibility overlay. Thirteen of 50 load a commercial overlay, the widget that adds a floating accessibility button. Four of the 12 clean homepages are among them. On one of them, the overlay itself put two invisible stops at the very start of the keyboard path.
Overlays change how a page looks in the browser; they don't fix the code that WCAG is measured against. They can also rewrite a page as it loads, which may make automated results look better than the site really is.
Smaller things that still matter: five homepages play a looping background video with no pause button, and two block pinch-to-zoom on phones.
Why "it passed the scan" isn't the answer
Automated tools are good at what can be measured from the code alone: missing alt attributes, contrast math, empty links. They can't tell whether focus is visible, whether a menu works from the keyboard, whether a skip link goes anywhere, or whether alt text actually describes the image. Those are exactly the things that lock people out.
It also means a content editor can do everything right and the site still fails. The fix belongs to whoever owns the template.
What I'd do before April 2027
- Find out who owns the template. For many public sites it's a vendor, a platform provider or an agency, not the people who edit the pages. Your compliance plan has to reach them, and it should be in the contract.
- Test by keyboard, not just by scanner. Unplug the mouse and press Tab through your homepage, your most-used form and anything interactive: menus, maps, sliders, quizzes. If you can't see where you are, or can't reach something, that's a failure.
- Fix the template-level issues first: icon labels, focus styles, brand colors, menus and video controls. They're cheap, and they clear the most problems at once.
- Don't buy an overlay as a shortcut. Spend that money on fixing the source.
- Write down what you checked and when. If you're asked about your progress, a dated record of testing and fixes is worth more than a scan screenshot.
- Don't forget documents. Most PDFs, forms and meeting packets you post are covered too (the rule has limited exceptions for archived and older documents), and they're often the biggest backlog.
About the method
I used axe-core 4.10.2, an open-source testing engine, against WCAG 2.1 A and AA rules on each homepage, plus a scripted keyboard walk-through on 16 sites with every finding confirmed by screenshot. Automated results are a starting point, not a verdict. This isn't legal advice, and the Title II date for each local government depends on its population.
The full report, with results by sector, five template fixes, a 15-minute keyboard self-test and sample contract language for web vendors, is at Had A Farm. If you manage a public website in Arkansas and want a second look before the deadline, that's what I built ComplyField and eiSEO to help with.