Manual Accessibility Audits Versus Automated Scans

  • Automated scans catch roughly a quarter of accessibility issues; the rest need human testing.
  • A (manual) audit evaluates context, keyboard flow, and screen reader behavior that scans cannot read.
  • Use scans to complement a (manual) audit, never to replace it.

A (manual) audit is conducted by a person who tests your site the way real users do, while an automated scan is a tool that flags code patterns it can detect. Scans are fast and cheap, but they only cover a fraction of what WCAG requires. Human testing carries the rest.

What each method actually catches

Automated scans read markup. They look for missing alt attributes, empty labels, and structural patterns a machine can measure. That work is real, but limited.

A (manual) audit goes where the machine stops. A person checks whether alt text is meaningful, whether a form makes sense with a keyboard, and whether a screen reader announces content in a logical order.

Here is how the two methods compare across the work a WCAG evaluation demands:

Manual (manual) audits versus automated scans by capability
Capability Automated Scan Manual Audit
Coverage of WCAG criteria Around 25% Full evaluation
Meaning of alt text No Yes
Keyboard-only navigation Partial Yes
Screen reader behavior No Yes
False positives Common Rare
Speed Seconds Days to weeks

Why scans miss so much

A scan cannot judge intent. It sees that an image has alt text, but it cannot tell you the alt text says “image123.jpg” instead of describing the photo.

The following are common issues that automated scans routinely miss:

  • Alt text that exists but describes nothing useful.
  • A tab order that jumps around the page illogically.
  • A custom dropdown that a keyboard user cannot open.
  • Error messages that never reach a screen reader.

These are the issues that stop real people from completing a purchase or reading a page. A machine reports the page as “clean” while users are stuck.

How the two work together

Scans are not useless. They are a starting point for practitioners, and we use them to complement human effort, never to replace it.

Here is the order we recommend for evaluating a site:

  1. Conduct a scan to surface quick, machine-detectable items.
  2. Conduct a (manual) audit across representative page templates.
  3. Remediate the confirmed issues, then re-test the fixes.

If you want a record of the results over time, keeping a running log of your evaluations helps you show progress and catch regressions after code changes.

What this means for WCAG conformance

WCAG conformance is a claim about how your site behaves for real users. A scan cannot make that claim on its own, because most success criteria require human judgment.

Put is a phrase we avoid, so here is the plain version: if a vendor sells you an “audit” that is really an automated report, you are paying for a slice and getting billed for the whole.

If it is automated, it is a scan. If it is an audit, it is manual. That distinction decides what your ACR can honestly say.

Getting an accurate audit

We only sell fully manual accessibility audits, and we deliver clear reports most clients can act on within one to two weeks. Our pricing stays competitive and our turnarounds are fast.

Want your cost and timeline? Send us a message and we will respond quickly, usually within a few hours.