Software can produce a score for your site, but that score reflects only what automated scans can detect, which is approximately 25 percent of WCAG issues. A conformance level score from software alone is not an accurate measure of WCAG 2.1 AA or 2.2 AA conformance. Real conformance requires a manual audit conducted by accessibility professionals, because the remaining 75 percent of success criteria require human evaluation, screen reader testing, and keyboard testing.
| Key Point | What It Means |
|---|---|
| Scan coverage | Software scans evaluate roughly 25 percent of WCAG success criteria automatically. |
| Score meaning | A software score reflects detected issues only, not full conformance status. |
| Conformance level | Levels A, AA, and AAA are established through manual audit findings, not scan scores. |
| Use of scoring | Scan scores work as monitoring signals across many pages, not as conformance verdicts. |
What Software Scoring Actually Measures
Accessibility scanning software loads pages and evaluates HTML, CSS, and ARIA attributes against a defined set of WCAG checks. The score is calculated from issues the scan can detect programmatically, such as missing alternative text, missing form labels, or invalid ARIA usage.
These are real issues, and surfacing them is valuable. The limitation is scope. A score of 95 means the scan identified few issues among the criteria it can check. It does not mean 95 percent of WCAG is satisfied.
Why Software Cannot Confirm a Conformance Level
WCAG conformance at Level A, AA, or AAA requires every applicable success criterion at that level to be satisfied. Many of those criteria require human judgment.
A scan cannot confirm whether alternative text accurately describes an image. A scan cannot verify that focus order matches a logical reading sequence. A scan cannot evaluate whether error messages help a screen reader user recover. These determinations come from manual evaluation, screen reader testing, and keyboard testing.
When software reports a conformance level, it is reporting against the subset of criteria it can check. A site can pass every automated check and still have significant issues at Level AA.
How Platforms Present Scoring Responsibly
Accessibility compliance platforms that combine scan data with audit findings present scoring in a more accurate way. The scan score reflects monitoring health across pages. The conformance status reflects the manual audit results, with each success criterion marked as Supports, Partially Supports, Does Not Support, or Not Applicable.
This separation matters. The scan score is a useful operational signal. The conformance level is an evaluation conclusion. Platforms that conflate the two produce reports that overstate compliance posture.
When Software Scoring Is Useful
Scoring has a legitimate place in an accessibility program when used for what it actually measures.
- Monitoring drift: scheduled scans flag when new issues appear after deployments.
- Page comparisons: scoring across hundreds of pages identifies which templates or sections need attention first.
- Pre-audit cleanup: fixing scan-detected issues before an audit reduces the issue volume an auditor documents.
- Trend reporting: scores over time show whether remediation work is reducing detected issues.
None of these uses produce a conformance verdict. They support the work that leads to conformance.
What to Look for in Platform Scoring Features
Evaluating how a platform presents scoring tells you how it treats accuracy. Platforms that present a single score as a measure of WCAG conformance overstate what their data supports. Platforms that distinguish scan results from audit results give an honest picture.
Look for features that show which success criteria the scan covers, which require manual evaluation, and how audit findings are integrated into the conformance view. Platform features that combine scan monitoring with audit-based conformance tracking give the most accurate read on where a site stands.
Software can score what it can see. Conformance is established by evaluating what it cannot.