The Work Isn’t Done When the Design Is Done
Ignoring accessibility laws for websites can lead to costly legal consequences, highlighting the importance of integrating compliance into design processes to avoid lawsuits and enhance user experience.
By Ray with my favorite human, Benjamin Scott. News Brief,
Let me catch you up. There's a batch of craft advice making the rounds this week, and it all points the same direction: the work isn't finished when the design is done. It's finished when someone else understands it, can use it, and won't sue you over it. None of this needs a new tool or a reorg. It's stuff your team can pick up this quarter.
The deep cut
- The deliverable is a product with users. MeasuringU ran its report template through four rounds of testing with real readers.
- Find the problems before you pay for users. Nielsen's ten heuristics catch a big share of issues with zero recruiting.
- Accessibility is a legal line, not a nice-to-have. Target paid roughly $6 million after saying the law didn't apply.
Your report is an interface too
You spend weeks on the research, then dump it into a slide deck nobody reads closely. Readers won't give your report the time you gave it. MeasuringU treats the report itself as an interface with two users: the researcher building it and the busy stakeholder skimming it.
Their nine fixes are basic craft. Clear structure, high-res images to break up text, two or three colors for hierarchy, left-aligned text, enough whitespace. Then the part people skip: usability test the template. Their current one went through four iterations with the research team. As they put it, if your report needs you in the room to explain it, it isn't finished.
The rule to steal: fix the weakest part of your template, then test it on someone who wasn't in the research.
Catch the problems before you recruit
You can find a large share of usability problems before a single user shows up. William Hudson calls expert evaluation one of the most cost-effective methods there is. No recruiting, no facilities, no scheduling. Just someone with judgment working through Nielsen's ten heuristics: system status, error prevention, recognition over recall, and the rest.
Two things worth noting. First, write your criteria as goals, not solutions. "Provide adequate navigational feedback" beats "provide breadcrumbs," because it stays true as patterns change. Second, on AI: it can draft a first list, but it works from a screenshot, misses the serious problems, and invents issues that aren't there. Treat it as a starting point, not the expert.
This doesn't replace user testing. It runs first, when recruiting would be premature.
The lawsuits are real and they add up
Accessibility isn't a moral bonus round. It's a legal risk with a price tag. Hudson lays out the history of accessibility litigation, and the numbers move fast. US federal website lawsuits jumped from 814 in 2017 to 2,258 in 2018, then past 4,000 a year now. Europe saw over 4,200 cases in 2023. The European Accessibility Act took effect in June 2025.
The cases set the precedent. Target thought the law didn't apply to its website. It cost about $6 million in legal fees and fixes. WCAG Level AA compliance is the common target, and it does double duty: it prevents complaints and shows you made reasonable efforts if someone sues. Building it in from the start beats retrofitting every time.
Small form choices carry the load
The place accessibility and usability meet is the form. Hudson's guidance on keyboard navigation is refreshingly concrete. Make the primary action loud and secondary links quiet, the way Amazon's "Place your order" button dwarfs the small links around it. Get tab order right, because screen readers only read the field label. Keep typers typing instead of forcing a mouse click into a dropdown.
Be generous with input. If someone types a card number with spaces, strip the spaces. One line of cleanup code beats a paragraph of error messages. There's a new reason to mark up forms properly: AI agents fill them using the same signals as screen readers. A form a screen reader can navigate is one an agent can complete.
Why people miss what's right in front of them
Even obvious changes go unseen. Hudson walks through three perceptual traps that wreck dynamic interfaces. The invisible gorilla study is the famous one: about half of viewers counting basketball passes never saw a person in a gorilla suit walk through. That's inattentional blindness, and it applies to your app.
The practical part is placement. He recounts a Barclays test where users kept missing a "next" button in the bottom-right corner, so the bank had to add text telling people to click it. On a shopping site, an "add to basket" counter updated in the top-right while users worked in the bottom-right, and nobody noticed. Ajax didn't cause that. Bad positioning did. Make updates big, colored, or animated, and put feedback where the eyes already are.
Make the insight land
The through-line under all of this: presentation decides whether the finding survives. Jane Fulton Suri says a real insight is an epiphany, not a summary, "a change of state in your head, and in the telling of that story in other people's heads." Tidy frameworks don't shift anyone. Stories do.
Two of her working habits are easy to adopt. Let conviction lead sometimes, without waiting for a data point to grant permission. Evidence and gut do different jobs. And treat research as co-discovery, two people curious about the same thing, not extraction with a clipboard. That's the same job MeasuringU's report design is doing, just earlier in the pipe.
Three questions for your team
- When did we last test our report template on someone who wasn't in the research? If never, that's this month's fix.
- Are we running a heuristic pass before we spend money recruiting users, or skipping straight to expensive testing?
- Where does our checkout or key form put its feedback, and can a screen reader and an AI agent both get through it?
