Operations and observability
Measure accessibility, don’t estimate it: a 32-page audit
Starting questionHow usable is a website really, for every user, in both color modes, and on every screen size?
Result
Works
Measured automatically, accessibility becomes a quality gate: contrast errors down by about 70 percent, language errors by about 62 percent.
With automated checks across every page, both color modes, and several screen widths: performance, accessibility, contrast, multilingual content, and layout. Figures before and after make progress provable; in our audit of 32 pages contrast errors fell by about 70 percent.
What we did
We tested 32 pages automatically in two color modes and four screen widths: performance, accessibility, contrast, multilingual content, and layout.
After the fixes, contrast errors dropped by around 70 percent and language errors by around 62 percent. Remaining work stayed open and was named.
What it means for business architecture
Accessibility is a requirement you can measure. Automating it repeatably turns it into a quality gate rather than a one-off exercise.
Lessons learned
- A light and a dark mode double the test scope. Errors often cluster in one of the two.
- Before-and-after measurements make progress provable.
- Open issues belong explicitly in the report, not swept under the rug.
- accessibility
- WCAG
- contrast
- quality assurance