Guide · Digital services
An accessibility governance maturity model
Published May 15, 2026 · 7 min read
A lightweight maturity ladder for public-sector and large-site programs: from ad-hoc audits to continuous operations with clear ownership, evidence, and procurement-ready reporting.
Accessibility programs rarely fail because teams lack good intentions. They fail because signal is noisy, ownership is unclear, and evidence does not survive contact with oversight. The ladder below is a conversation aid — not a certification scheme — for where your organization sits today and what to fund next.
Level 1 — Ad hoc
Periodic manual audits or one-off scans before launch. Findings live in PDFs or tickets with no durable link back to templates or components. Regression risk is high because nobody owns rescans after release.
Level 2 — Project-led
Defined owners for major releases; some reuse of test cases. Tooling may run in CI or before go-live, but prioritization is still page-by-page and cross-team coordination is brittle.
Level 3 — Continuous detection
Scheduled crawling across the real site inventory, alerts on regressions, and a shared view of severity. Automated coverage is understood as partial — manual review is budgeted for high-traffic templates and flows.
Level 4 — Workflow and accountability
Findings deduplicated into remediation units tied to code or CMS objects; assignments, due dates, and verification runs exist. Legal and procurement can trace what was open, what changed, and when.
Level 5 — Program integration
Accessibility criteria embedded in procurement, vendor onboarding, and content governance. Reporting rolls up to portfolio dashboards; exception paths are documented; VPAT or equivalent evidence is maintained as a living artifact, not a annual PDF refresh.
Using the model without vanity metrics
Maturity is not “how many issues you closed this quarter.” It is whether the organization can answer three questions on demand: What do we publish? What is failing WCAG today? What did we change since the last oversight conversation? Tools like Parity support levels 3–5 when paired with executive sponsorship and manual review where automation ends.