How-To Guides
One scan vs. continuous monitoring: which actually keeps you compliant?
Ronny Teichgräber
Founder, Accessive
A one-time WCAG scan tells you the score on the day you ran it. Continuous monitoring tells you the score every day after. For US teams under ADA pressure, only one of those is a compliance posture.
A one-off accessibility scan is useful — it gives you a snapshot of where your site stands today against WCAG 2.2 AA, ADA Title III and Section 508. But that snapshot starts going stale the moment the next deploy ships. For US teams trying to defend against demand letters, the question is not "did we pass once" — it's "are we still passing today, and can we prove it for any day in the last 90?"
What a single scan tells you
A scanner like axe-core or the Accessive scanner enumerates the WCAG checks it can automate, flags violations, and gives you a score. The output is excellent — but it is a single observation in time. The moment a designer ships a new hero, a developer adds a new form, or a marketer drops in a third-party widget, the report you ran last week no longer describes your live site.
What continuous monitoring tells you
Continuous monitoring re-runs that same enumeration on a schedule (daily for most teams), keeps the time-series, and surfaces changes: new violations introduced by recent deploys, regressions on pages that used to pass, and the audit trail that shows when each issue first appeared and when it was resolved.
For ADA defense, the audit trail is the asset. A demand letter from a
serial plaintiff firm typically arrives weeks or months after the
issue was introduced. Without a time-stamped record of when an issue
appeared and when you fixed it, your engineering team is reconstructing
history from git log and guesswork.
When a single scan is enough
If your site is genuinely static — a docs page that hasn't changed in two years, a legal-page that ships once a quarter — a one-off scan is fine. Run it, fix what it finds, archive the report.
When you need monitoring
If you ship more than once a week, embed third-party widgets, run A/B tests on real users, or have any kind of CMS where non-engineers can publish, a snapshot is structurally insufficient. The question is not whether you'll regress — it's whether you'll know.
