CASE STUDY: ACCESSIBILITY CONFORMANCE REPORT (ACR)
VPAT® 2.5Rev WCAG Edition | Corporate Marketing Website
WCAG 2.0, 2.1 & 2.2 Level A and AA | May 2026
Project Context:
Client: Content development firm (name redacted) | Public-facing corporate marketing website
Role: Accessibility Specialist, LIFTUPS™ Digital Consulting
Deliverable: Accessibility Conformance Report (ACR) based on VPAT® Version 2.5Rev, WCAG Edition
Engagement: Scoped WCAG evaluation of four representative pages (Our People, Case Studies, Content Development services, and Accessibility Statement), delivered as a formal ACR
Tools: WAVE (WebAIM), axe DevTools Pro (axe-core 4.11.4), Colour Contrast Analyser (TPGi), NVDA, JAWS
Standards: WCAG 2.0, 2.1, and 2.2 Level A and AA
Download the full Accessibility Conformance Report (PDF, opens in a new tab)
About this report: To protect client confidentiality, the client’s name and domain are redacted in the report and on this page.
Challenge
An Accessibility Conformance Report is how an organization tells buyers, including federal agencies reviewing products under Section 508, how well its product meets accessibility standards. Each WCAG success criterion gets a conformance level and an explanation. Procurement teams make decisions from that document, so every conformance level has to be backed by test results, not assumptions.
The client needed its public corporate marketing website evaluated against WCAG 2.0, 2.1, and 2.2 Level A and AA, with the results documented in the industry-standard VPAT® 2.5Rev format. The site used CSS gradients, background images, a three-column staff grid, and a third-party AI chat widget that loaded after the page. Each of those is a place where automated scanners are known to miss real problems or report problems that are not there.
Approach: Multi-Tool Testing With Manual Verification
No single accessibility tool gives a complete picture. I ran two automated scanners that work in different ways, then used manual tools to confirm which flagged issues were real failures and which were false positives:
- WAVE by WebAIM: Scans the page as it is visually rendered, catching contrast, structure, ARIA, and form label issues
- axe DevTools Pro (axe-core 4.11.4): Scans the underlying code (DOM) against WCAG 2.1 AA and WCAG 2.2 AA rules
- Colour Contrast Analyser (TPGi): Manual contrast ratio checks using the actual CSS color values from browser DevTools
- NVDA and JAWS: Manual screen reader testing of reading order, focus behavior, and how interactive components are announced
- Keyboard-only testing: Tab order, skip link behavior, and visible focus
Why the Scanners Disagreed, and How I Resolved It
The most important part of this evaluation was the gap between what the tools reported. Three patterns showed why manual cross-checking is required before any conformance level goes into an ACR:
- Contrast: WAVE flagged 98 contrast errors; axe flagged 0. axe reads color values from the code and cannot calculate contrast for text sitting on gradients, background images, or transparent containers, so it reports nothing rather than guess. WAVE evaluates the rendered page and flagged errors on every page. Manual checks with the Colour Contrast Analyser separated the real failures from the false positives.
- Chat widget: WAVE caught it; axe missed it. The third-party AI chat widget loads through a JavaScript bundle after the page finishes loading, outside the snapshot axe evaluates. WAVE, which tests the fully rendered page, found its missing form label.
- Empty links: axe caught them; WAVE missed them. On the Case Studies page, axe flagged 11 serious “links must have discernible text” errors. WAVE reported 0 errors on the same page and gave it the highest score of the four pages (7.7).
Manual contrast verification (Colour Contrast Analyser):
| Colors Tested | Measured Ratio | Result | Conclusion |
|---|---|---|---|
| #B1B8D9 on #15225A | 7.6:1 | Pass | False positive |
| #E8E6E6 on #616992 | 4.3:1 | Fail | Confirmed failure |
| #5896C6 on #15225A | 4.7:1 | Pass | False positive |
| Navigation element | 4.28:1 | Fail | Confirmed failure |
What the Evaluation Found
WAVE AIM scores and contrast errors by page:
| Page | WAVE AIM Score (0–10) | Contrast Errors Flagged |
|---|---|---|
| Our People | 4.4 | 43 |
| Content Development | 5.6 | 33 |
| Accessibility Statement | 6.8 | 11 |
| Case Studies | 7.7 | 11 |
Key findings:
- Unlabeled links (2.4.4, Does Not Support): 11 links on the Case Studies page had no accessible text. A screen reader user pulling up the page’s list of links would hear 11 links in a row with no indication of where any of them lead.
- A chat widget that says “false” (4.1.2): The chat widget’s aria-label read
"Chat with AI, false, ". A true/false value had been passed into the label as literal text, so a screen reader would announce “Chat with AI, false, comma.” The widget’s input field was also missing a label entirely. - Reading order that does not match the layout (1.3.2): The staff grid displays profiles in rows, but the code reads them column by column. NVDA testing confirmed a screen reader user hears profiles in a different order than sighted users see them.
- Alternative text (1.1.1): 13 staff photos had suspicious or inadequate alt text, and other pages had empty or questionable alt text needing review.
- Structure (1.3.1, 2.4.6): Skipped heading levels confirmed by both tools, text styled as headings without heading markup, and form groups missing their legends.
- The Accessibility Statement page itself (1.4.3): The page making public accessibility claims scored 6.8, with 11 contrast errors and 54 ARIA instances flagged, so I recommended it for priority remediation.
Conformance summary:
| Conformance Level | Level A Criteria | Level AA Criteria |
|---|---|---|
| Supports | 13 | 9 |
| Partially Supports | 7 | 6 |
| Does Not Support | 1 | 0 |
| Not Applicable | 10 | 5 |
| Not Evaluated (outside evaluation scope) | 1 | 4 |
Five criteria that need interactive form and target testing (3.3.1, 3.3.3, 1.3.5, 2.4.11, and 2.5.8) were reported as Not Evaluated, with the reason documented in the report. I did not assign a conformance level that the testing could not support.
Recommendations Delivered
- Start contrast remediation on the Our People page (43 instances, lowest score), beginning with the two confirmed failures at 4.3:1 and 4.28:1
- Give all 11 Case Studies links descriptive accessible names, using visible text, aria-label, or the related case study title
- Work with the chat widget vendor to add a label to the input field and correct the aria-label to a clean value such as “Chat support”
- Reorder the staff grid markup so the code order matches the visual row order (WCAG technique C27)
- Write descriptive alt text for informational images, including each staff member’s name and role, and mark decorative images with empty alt text
- Fix skipped heading levels, mark up visual headings as real headings, and add legends to form groups
Results & What This Demonstrates
- Defensible conformance levels: Every Supports, Partially Supports, and Does Not Support rating is backed by tool results and manual verification
- False positives separated from real failures: Manual contrast checks showed that two of the four flagged color pairs actually passed, so remediation effort goes where it is needed
- Issues one tool alone would have missed: The empty links, the chat widget label, and the column-first reading order would each have been missed or misreported by at least one scanner
- Actionable for developers: Each finding names the page, the WCAG criterion, the number of instances, and a specific fix
Compliance Standards
- WCAG 2.0, 2.1, and 2.2 Level A and AA (Web Content Accessibility Guidelines)
- VPAT® Version 2.5Rev, WCAG Edition (Information Technology Industry Council)
- Section 508 (federal accessibility standard, which incorporates WCAG 2.0 Level A and AA)
Tools & Technologies
Automated Testing: WAVE by WebAIM v3.3.1.0, axe DevTools Pro (axe-core 4.11.4)
Manual Verification: Colour Contrast Analyser (TPGi) v3.5.5, browser DevTools, keyboard-only testing
Screen Readers: NVDA 2025.3.2, JAWS 2026
Reporting: VPAT® 2.5Rev WCAG Edition
Reflection
An ACR is only as useful as it is accurate. A report built only on automated scans would have rated the Case Studies page as clean, missed the chat widget’s broken label, and reported dozens of contrast errors that turned out to be false positives. Any of those mistakes would mislead the people relying on the report: the buyer deciding whether to purchase, and the users with disabilities who depend on the product working.
The value of an accessibility specialist is not running the tools. It is knowing where each tool is blind, testing those gaps by hand, and writing conformance levels that a procurement team, a developer, and a screen reader user can all trust.
Need an Accessibility Conformance Report for your product? Contact me
