RGAA / accessibility
A pragmatic path to RGAA/WCAG compliance: a prioritised action plan by effort/impact, a conformance score, and a day-rate estimate.
Accessibility is not a checkbox after delivery: it is a managed workstream with a starting score, a client target, and a costed path to get there. Our method is built on an action plan grouped by effort/impact, not an unordered list of non-conformant criteria.
Prioritised action plan
Three levels: P1 quick wins (low effort, high impact, often a single cross-cutting change), P2 structural fixes (shared components - one fix corrects multiple pages), P3 complementary fixes (needed to reach 100%). Each priority comes with a criteria sub-total and an estimated score.
Day-rate estimate
Each defect is estimated individually (very low / low / medium / high), then converted into person-days broken down by profile: Front, Back, Content. The CIO and the client know exactly what each phase costs before signing off.
Measurable conformance
The RGAA (or WCAG) conformance score is calculated from the audit and updated after each remediation batch. The 75% intermediate target is often reachable with P1+P2 alone, delivering immediate ROI on the remediation budget.
The RGAA framework
RGAA (the French General Accessibility Improvement Reference, maintained by DINUM) describes 106 criteria across 13 themes. It transposes WCAG 2.1 Level AA into French law. Conformance is mandatory for public services and large companies (revenue above 250 M EUR), and increasingly required in private-sector tenders.
Our method: the effort/impact action plan
After receiving an audit report (from IPEDIS, access42 or another accredited body), we do not hand back the raw list of non-conformant criteria. We produce a structured action plan in three priorities:
Priority 1 - Quick wins (low effort, high impact)
Cross-cutting fixes that apply to every page with a single code change: lang attribute on <html>, background-color on <body> for contrast, alt="" on decorative images, skip link, :focus-visible outline. Typically 2 to 3 Front person-days, immediately visible in the score.
Priority 2 - Structural fixes (medium effort, high impact)
Shared components: modals (focus trap, aria-modal, aria-labelledby, keyboard close), main navigation (role="navigation", aria-expanded), forms (<label> for/id, accessible error messages), contrast on reused components. A single fix corrects multiple pages and multiple criteria simultaneously.
Priority 3 - Complementary fixes (variable effort, required for 100%) Video subtitles, descriptions for complex images and infographics, accessible PDFs, editorial content audits. These items often fall partially on the client side (editorial content, CMS admin) rather than the agency.
Person-day estimate
Each defect sheet includes a qualitative estimate (very low = 1 CSS line; medium = touches several React components). These are converted into person-days and presented in a summary table:
| Phase | Target score | Front (d) | Back (d) | Content (d) | Total |
|---|---|---|---|---|---|
| P1 Quick wins | ~63% | 1.5 | 0.5 | - | 2 |
| P2 Structural | ~75% | 7 | 0.5 | - | 7.5 |
| P3 Complementary | 100% | 6 | 1 | 3 | 10 |
Figures vary by project. The example above is a ballpark based on real projects (BPCE Recrutement, LFB Website).
What this means for your CIO
- No audit without an operational follow-up: the action plan is delivered alongside the corresponding merge requests and a re-audit strategy.
- Full traceability: every defect is linked to a numbered RGAA criterion, a failing page, and a remediation MR.
- Code vs content split: code fixes are delivered by the agency; editorial fixes (image alt text, video subtitles) are documented for the client with an implementation guide for the CMS.
- Scheduled re-audit: we annotate the auditor's grid before resubmission to ensure corrections are correctly assessed as conformant.