What happens when a vendor hands you a glossy “98% accessible” score, your legal team asks how that number was produced, and nobody can answer? That scenario plays out more often than most organizations expect. Picking an auditor is not a shopping decision about dashboards. It is a decision about evidence, and evidence has to survive scrutiny from lawyers, procurement officers, and actual disabled users.
Quick Answer: How do I choose a web accessibility auditing company?
Choose a web accessibility auditing company by demanding proof of method before you discuss price. A credible vendor audits against WCAG 2.2 at a stated conformance level, tests manually with assistive technology, and names the people who do the work.
Some firms sell one-off conformance reports, others bundle remediation and continuous monitoring, and a few sell overlay widgets that claim to fix everything automatically. Those are very different products, and only the first two produce defensible evidence.
Before signing, compare WCAG version and target level, manual versus automated testing mix, tester credentials, report format and re-testing, independence from overlay products, sample deliverables, and total cost including remediation support.
Why the Auditor Choice Matters More Than the Report

Accessibility failures are rarely dramatic. They are quiet: a form field with no label, a modal that traps keyboard focus, a carousel that cannot be paused. None of these break a page for most visitors, which is exactly why they survive internal QA. An auditor’s job is to find what your own team cannot see, and the quality of that finding depends entirely on who is looking and how.
The stakes have also shifted from reputational to legal. The European Accessibility Act (Directive (EU) 2019/882) applies to many products and services from 28 June 2025. In the United States, the Department of Justice’s April 2024 final rule under ADA Title II sets WCAG 2.1 AA as the standard for state and local government web content, with compliance dates running from April 2026 for larger entities to April 2027 for smaller ones. If your organization sits anywhere near those rules, an audit is no longer a nice-to-have document. It is potential evidence.
That changes the selection criteria. You are not buying a score. You are buying a defensible record of what was tested, how, and by whom.
Start With the Standard, Not the Sales Deck
The first filter is simple and eliminates a surprising number of vendors. Ask which version of WCAG they audit against and which conformance level they target.
WCAG 2.2 became a W3C Recommendation on 5 October 2023. It adds nine success criteria to 2.1 and removes 4.1.1 Parsing. WCAG 2.1 dates from 5 June 2018, and WCAG 2.0 from 11 December 2008. A vendor still auditing only to 2.0 is working from a baseline that predates most modern web patterns, including the mobile and touch interactions your users rely on.
Level matters just as much. Level A is the floor, AA is the level most laws and policies reference, and AAA is rarely achievable across an entire site. A vendor that promises full AAA conformance on a large site is either misunderstanding the standard or telling you what you want to hear.
There is a practical angle here too. If you already run security tooling and know how carefully vendors describe their coverage, you will recognize the pattern. The same discipline you would apply to choosing a web application firewall that fits your stack applies to auditors: vague coverage claims are a warning sign, and specifics are the whole game.
Manual Testing Is Not Optional, and Automation Alone Cannot Save You
Automated scanners are useful. They catch missing alt attributes, low contrast in some cases, and structural problems fast. They are also blind to a large share of WCAG failures, because many criteria depend on meaning and context that no script can evaluate.
Consider a few examples. Whether alt text gives the right information depends on a human who understands the image’s purpose on that page. Whether keyboard focus order makes sense depends on someone tabbing through the flow. Whether a component creates excessive cognitive load depends on judgment, often from people who use the interface differently than your designers do.
Industry studies consistently find that automated testing catches only a minority of WCAG failures, though the exact percentage varies widely by study and methodology. That variance is worth noting: if a vendor quotes a precise automation figure, ask which study and which test corpus produced it.
The practical takeaway is blunt. Treat “100% automated” as a disqualifier, not a feature. If you want the underlying mechanics of how scanning tools intercept and block requests, the explainer on how a web application firewall works for beginners covers similar detection logic, and it makes clear why pattern matching has limits.
Questions That Separate Serious Auditors From Poseurs
Bring a fixed list to every sales call. The answers, and the comfort with which vendors give them, tell you more than any case study.
| Question | Weak answer | Strong answer |
|---|---|---|
| Which WCAG version and level? | “We cover all of them” | Names 2.2 AA and explains why |
| Who performs the testing? | “Our platform” | Named testers, some with disabilities |
| Manual testing included? | “Our AI handles it” | Manual review plus assistive tech testing |
| What assistive tech is used? | Not specified | Screen readers, magnifiers, switch access, voice input |
| What does the report contain? | A score out of 100 | Issue-by-issue mapping to success criteria |
| Is re-testing included? | Costs extra, undefined | Defined re-test after remediation |
Notice that none of these questions are about price. Price comes later, once you know what you are comparing.
Ask Who Actually Does the Testing
Credentials are a reasonable proxy for rigor. IAAP certifications such as CPACC and WAS (Web Accessibility Specialist) signal that a tester has passed a structured exam on accessibility practice. They are not a guarantee of quality, but an absence of any credential across an entire team is worth questioning.
Equally important is whether people with disabilities are involved in testing. An auditor who has never watched a screen reader user attempt a checkout flow is describing theory. Real sessions surface problems that checklists miss, particularly around timing, error recovery, and navigation.
Demand the Methodology in Writing
A credible audit states the WCAG version and level, the pages and user journeys sampled, the browsers and assistive technologies used, and whether findings are reproducible. If a vendor cannot describe sampling logic, you have no way to know whether they tested your three most important flows or three pages that happened to be easy.
Ask for a redacted sample report. Most established firms will share one. Read it closely for reproduction steps. A finding that says “contrast issue on homepage” is nearly useless. A finding that names the element, the measured ratio, the required ratio, and the success criterion is something your developers can act on and your lawyers can defend.
Audit, Remediation, or Monitoring: Know Which You Need
Vendors differ sharply in scope, and mismatched expectations cause most of the friction in this market.
A one-off conformance audit produces a point-in-time report. It suits procurement requirements, pre-launch checks, and legal defensibility.
Remediation support means the vendor helps fix what it found, sometimes with developer pairing or code-level guidance. This is where most organizations actually need help, because a report nobody can implement changes nothing.
Continuous monitoring adds ongoing scanning and periodic manual review. It fits sites that ship changes weekly, where a six-month-old report describes a product that no longer exists.
Decide which of these you need before you request quotes. Otherwise you will compare a monitoring subscription against a full audit and conclude, wrongly, that one vendor is four times cheaper. The same scoping trap appears in other procurement categories: the guide on choosing privileged access management software that fits your needs makes a similar point about matching tool scope to actual risk rather than to a feature list.
Watch for Conflicts of Interest
Here is a pattern worth flagging plainly. A vendor that sells an accessibility overlay widget and also performs audits has a structural incentive to find that its own widget solved most problems.
Overlays are contested. Disability advocates and researchers have widely criticized widgets that claim to automatically “fix” sites, arguing they do not achieve conformance and can interfere with the assistive technology users already rely on. Whatever position a buyer takes, an auditor grading its own remediation is not independent, and independence is the entire value of an audit.
Where conformance claims matter for legal or procurement reasons, prefer an auditor with no product to sell you on the side.
Practical Tips From the Field
A few habits make the selection process smoother, and a few mistakes recur often enough to name.
Test a small paid pilot first. Ask a shortlisted vendor to audit one critical user path before you commit to a full engagement. Compare the pilot reports side by side. The difference in depth is usually clear within an hour of reading.
Avoid the lowest-bid trap. Very cheap audits are usually automated scans with a template attached. You will pay again when the report fails to satisfy a regulator or an internal accessibility committee.
Do not let the audit become a one-time event. The most common failure is treating the report as the finish line. Assign owners for each finding and set a re-test date before the engagement ends.
Bring developers into the kickoff. Auditors who understand your stack produce more actionable findings, and developers who hear the methodology directly are far more likely to accept the results.
Check the deliverable format early. A spreadsheet with 400 unprioritized rows is technically thorough and practically useless. Ask how severity is ranked and whether fixes are grouped by component so one code change resolves many instances.
If your team already evaluates tools by fit rather than by marketing, the instinct carries over. The same approach used in choosing AI transcription tools that actually fit your workflow works here: define the job, then find the tool that does that job with evidence.
What Fair Pricing and Timelines Look Like
Costs vary enormously by scope, so treat any single figure with suspicion. What you can compare is structure.
Expect pricing to reflect page count or user path count, the number of templates audited, whether assistive technology testing is included, and whether remediation support or re-testing is bundled. A mid-size site with a handful of templates and a defined set of critical paths is a different engagement from a large e-commerce platform with hundreds of components.
Timelines follow the same logic. A focused audit of core user paths can finish in a few weeks. A full-site engagement with remediation cycles runs into months. Be wary of a vendor promising a full manual audit of a large site in a week; that timeline usually means automation with a manual skim.
Ask explicitly what happens after the report lands. The best vendors schedule the re-test at signing, because they know the report is only useful if someone verifies the fixes.
Conclusion
Choosing a web accessibility auditing company comes down to one habit: asking for evidence and reading the answer carefully. Confirm the WCAG version and level, insist on manual testing with assistive technology, check who performs the work and what credentials they hold, review a sample report for reproduction steps, and keep the auditor independent from any overlay product. Match the scope to what you actually need, whether that is a one-off conformance report, hands-on remediation help, or continuous monitoring. Do those things and the decision becomes far less about persuasive sales decks and far more about whether the deliverable will hold up when someone finally asks how the number was produced.
How long does a professional web accessibility audit take?
Most focused audits of core user paths take a few weeks, while full-site engagements with remediation cycles run for months. The timeline depends on page and template count, how many assistive technologies are tested, and whether re-testing is included. A vendor promising a full manual audit of a large site within days is almost certainly relying on automation alone.
Is an automated accessibility scanner enough to prove conformance?
No. Automated tools catch only a minority of WCAG failures, and many criteria depend on human judgment about meaning and context. Appropriate alt text, logical keyboard order, and cognitive load cannot be evaluated by a script. Scanners are useful for fast structural checks, but conformance claims need manual testing, ideally including users of assistive technology.
Should I hire a company that also sells an accessibility overlay?
Be cautious. A vendor that audits sites and also sells an overlay widget has an incentive to find that its own product resolved most issues. Overlays are widely criticized by disability advocates and researchers as not achieving conformance. Where conformance claims need to withstand legal or procurement scrutiny, an independent auditor with no product to sell is the safer choice.
What should a usable accessibility report actually contain?
Each issue should map to a specific WCAG success criterion and include severity, the exact location, and reproduction steps a developer can follow. The report should state the WCAG version and level, the pages and journeys sampled, and the browsers and assistive technologies used. A single aggregate score is not conformance evidence and will not satisfy a regulator or an internal review board.
Do IAAP certifications actually matter when picking an auditor?
They are a useful signal rather than a guarantee. IAAP credentials such as CPACC and WAS show that a tester has passed a structured exam on accessibility practice. What matters just as much is whether people with disabilities take part in testing, because real sessions surface problems that checklists miss, especially around timing, error recovery, and navigation.
How do I compare quotes from different auditing companies fairly?
Fix your scope before requesting quotes. Decide whether you need a one-off conformance report, remediation support, or continuous monitoring, then ask every vendor to price the same pages, journeys, and testing methods. Otherwise a monitoring subscription will look far cheaper than a full audit when the two are not comparable products at all.


