Question 1

What does 'strict' mean vs 'moderate'?

Accepted Answer

Strict = axe-core's critical + serious impact ratings. These map to actual WCAG 2.1 AA failures — the ones that legally and ethically matter. Moderate covers advisory rules (region landmarks, heading-order suggestions) that are technically failures but rarely block keyboard or screen-reader users.

Question 2

How is the scan run?

Accepted Answer

Daily via Playwright + chromium against 10 representative URLs covering each page template (home, lane, methodology, insight article, case study, report, onboarding walkthrough, pricing, contact, 404). Each URL × two viewports (1440×900 desktop + 390×844 mobile). Scan duration ~2 minutes; results land at 06:00 local time.

Question 3

Why only 10 URLs and not every page?

Accepted Answer

Every page on the site uses one of the 10 templates. A clean scan on the templates means a clean scan on every page using that template. If we changed a template and didn't catch it in the daily scan, an axe-core rule would fail across all sibling pages — the 10-URL sample is the canonical regression set.

Question 4

What if my own scan turns up different numbers?

Accepted Answer

Likely cause: different tag set (we use wcag2a + wcag2aa + wcag21a + wcag21aa + best-practice) or different viewport sizes. The exact configuration is documented in scripts/axe-scan.mjs in the repo. Match those settings to reproduce the published numbers.

Question 5

How would I do this for my own site?

Accepted Answer

The script is 80 lines of vanilla Playwright + @axe-core/playwright in the public repo. Fork it, point it at your URLs, run it on a cron. For ongoing engineering follow-through (fixing the regressions that pop up week-to-week), that's exactly what the Audit Retainer lane delivers.

Accessibility report — live.

What “strict” means

Violations by rule

By page