A real audit is two layers stacked together — an automated scan plus a person actually using your site with a keyboard and a screen reader — and it produces a punch list, not a badge.
A real website accessibility audit tests your site against WCAG 2.0 Level AA using two layers stacked on top of each other: an automated scan that checks what a machine can verify, and manual testing that checks what only a human, using a keyboard, a screen reader, or both, can actually confirm.
The automated layer runs tools like Lighthouse, axe, or WAVE against your markup. They’re fast, consistent, and genuinely good at a narrow set of problems: missing alt attributes, contrast ratios that fail the math, form fields with no linked label, a handful of obvious ARIA mistakes. Feed one a URL and you’ll have a report in minutes.
What automated tools can’t do is use your site the way a disabled visitor actually would. An automated scanner catches roughly 30 to 40 percent of WCAG violations on a typical page, according to accessibility-testing platform TestParty’s breakdown of automated coverage — the other 60 to 70 percent only turn up when someone unplugs the mouse and tabs through the page, or turns on VoiceOver or NVDA and listens to what actually gets announced. Alt text that’s technically present but reads “image047.jpg” passes an automated scan and fails a human in about two seconds. So does a custom dropdown that looks fine and traps keyboard focus the moment you try to leave it.
That two-layer combination, scan first, then a person actually operates the page, is what separates an audit from a scan. The eight failure types both layers turn up constantly are broken down in full, with fix examples, in the WCAG 2.0 AA guide; this piece is about the audit process itself, not the individual fixes.
Three things get called an “accessibility audit,” and they are not the same product. Line them up and the gaps are obvious.
| Factor | Automated scan alone | Overlay/widget “scan” | Full audit (automated + manual) |
|---|---|---|---|
| What it actually tests | Programmatic DOM checks: alt attributes present, contrast math, some ARIA patterns | The same automated layer, usually a licensed scanning engine, wrapped in a branded dashboard | Everything automated catches, plus a real keyboard-only pass and a screen reader test |
| WCAG coverage | Roughly 30–40% of success criteria | Same automated ceiling — it’s still a machine underneath | Full success-criteria coverage, tested the way an actual visitor uses the page |
| What you get back | A technical list of flagged elements | A “score” or badge, often overstated | A prioritized, plain-English punch list mapped to WCAG criteria, with fixes |
| Confirms WCAG/AODA conformance | No — flags problems, can’t confirm a pass | No — the FTC fined one vendor $1,000,000 in 2025 for implying otherwise | The closest you get to a defensible answer, once the fixes actually ship |
| Typical cost | Free to a few hundred dollars | Recurring subscription, often $50+/month | Scales with site size — see the cost breakdown below |
Notice the middle column. An overlay’s dashboard-facing “scan” is still just an automated tool under the hood, running the same class of checks as a free scanner, dressed up as something more authoritative. More on exactly why that distinction matters further down.
You get a written, prioritized punch list, not a single pass/fail score and not a certificate. A real audit report ties every finding to a specific WCAG 2.0 AA success criterion, describes the problem in plain English, shows you exactly where it lives on the page, and tells you what to do about it.
Reports typically group findings by severity, because not every failure carries the same weight. A keyboard trap inside your main navigation blocks a task completely, that’s critical. A vague “click here” link is real friction, but a keyboard user can usually still get where they’re going, that’s lower priority. Ranking by actual user impact, not just by how many issues a scanner flagged, is what turns a report into something you can act on instead of a wall of 200 line items nobody reads past page two.
My own free scan is an automated pass on your key pages, returned as a plain-English punch list rather than a raw export from a scanning tool.
A free scan is automated and comes back in minutes; it covers your key pages, not your entire site.
A full audit takes longer, and how much longer depends on how many genuinely different templates and interactive components you’re running, not your page count. A ten-page brochure site built from three templates audits faster than a five-page site carrying a custom booking widget, a mega menu, and an embedded video library, because the manual testing scales with distinct components, not URLs.
Whatever the timeline, the manual portion is the part that can’t be compressed by throwing more automated tools at it. A person still has to sit down, unplug the mouse, and go through the site the way a keyboard-only or screen-reader user actually would.
Published pricing for a professional website accessibility audit generally lands between $1,500 and $5,000 for a small business site, per DigitalA11Y’s 2026 audit cost guide, with ADA Compliance Pros’ pricing breakdown putting small-business engagements at $3,000 to $5,000 and noting the wider market runs anywhere from $500 to $10,000 depending on scope.
Some vendors price per page instead of per project. ADA Compliance Pros’ published model charges roughly $100 for a static marketing page, up to $250 for a complex page carrying e-commerce logic or custom widgets, and $25 to $100 for light content like a policy page, because, as that guide puts it, page count alone doesn’t tell you much. A checkout flow with custom components takes several times the manual effort of a static policy page, even though both count as “one page.” Expert manual audits scoped for up to about 12 pages commonly run $2,500 to $10,000, and complex web applications can push past $25,000.
I quote my own audit and remediation work individually, after scoping the actual site. The size of the real number depends on what the audit turns up, not a flat rate off a price sheet.
One number worth sitting with regardless of which vendor you use: whatever you’d pay monthly for an overlay subscription adds up, over a year or two, to roughly the same range as a real one-time audit and fix, except the audit actually produces something that holds up. Why an overlay doesn’t get you there goes deeper on that math.
Read it by severity first, not by WCAG clause number. A well-built report ranks findings critical, high, medium, low, by real-world impact, so you can see immediately which handful of fixes unblock the most people instead of drowning in 150 line items sorted alphabetically by success criterion.
Each finding should carry four things: the specific WCAG success criterion it fails, so you know exactly what standard you’re being held to, a plain-English description of the actual problem, precisely where it lives (a template, a component, a URL), and a concrete fix, not just “improve contrast.” “Improve contrast” isn’t actionable. “Change the button text from #767676 to #595959 on a white background to clear the 4.5:1 ratio” is.
Two more things worth checking before you file a report away. Does it distinguish between a true failure and a judgment call — some WCAG criteria genuinely have room for interpretation, and a report that pretends everything is black-and-white is overclaiming. And does it include a re-test step. A report that ends at “here’s what’s wrong” without a path back to “here’s confirmation it’s fixed” leaves you exactly where a compliance report needs you not to be: attesting to something nobody actually verified.
No. An overlay’s built-in “scan” is the same automated layer described above, run by the vendor whose subscription you’re paying for, and it has the same ceiling every automated tool has: it can flag a problem, it can’t confirm your site passes.
The distinction matters because overlay vendors have specifically marketed their scan-and-patch combination as a substitute for real compliance work, and regulators have pushed back hard on that claim. The full story, including the FTC’s $1,000,000 order against one major overlay vendor and the 2025 lawsuit data, is in does an accessibility overlay make your website AODA compliant? Short version: no, and an overlay’s own scan report isn’t evidence otherwise.
Anyone who wants an honest answer needs one eventually, but the trigger that actually forces the question is your Ontario employee count. At 50 or more employees, a real audit is effectively mandatory groundwork — you can’t attest to WCAG 2.0 AA conformance on your compliance report without one, and that report is due December 31, 2026.
Below 50, the IASR’s website mandate doesn’t legally reach you, but the audit is still worth having. A 20-to-49-employee organization files the compliance report either way; knowing what an audit would find means you’re not guessing when you check that box. Employee count has never been a defence against a Human Rights Code complaint or a locked-out customer walking away. The full sector and threshold picture is in accessibility compliance in Ontario and the 20-vs-50 employee breakdown.
Run the self-checks yourself first; bring in a professional audit when the stakes or the complexity outgrow what a spot check can responsibly cover. Both are legitimate, depending on what you actually need out of the result.
You can genuinely test a lot yourself this afternoon: tab through your homepage with no mouse, run a free contrast checker, try your top few pages with VoiceOver or a browser screen-reader extension. That’s real signal, and it’s free. Where it stops being enough is the moment you need a defensible, documented record, for a compliance report, for a legal question, or because the site is big or interactive enough that a spot check would miss whole categories of components.
That’s the point where I’d rather do the full pass myself than hand you a checklist and hope. On any serious build, accessibility gets tested with the same rigor regardless of the client’s size; the difference at a smaller business isn’t the standard, it’s the scope. If you want a second opinion on where your site actually stands, the accessibility service covers exactly this, and a free scan is one message away at contact.
Start with the two-layer test above on your own site this week: automated scan, then a five-minute keyboard tab-through of your homepage. If it turns up more than you expected, that’s normal. Most sites built on a stock theme were never tested this way.
From there, go deeper wherever your business actually sits. The full legal picture, thresholds, the deadline, penalties, is in the AODA website accessibility guide. The Ontario-specific compliance landscape, including how it compares to federal law, is in accessibility compliance in Ontario. And if what you really need is someone to do the full audit and the fix, that’s what I do.
Quick, direct answers to what business owners ask most about website accessibility audits.
Audit, remediation and the compliance report, handled end to end for Ontario businesses.
AODA & WCAG Accessibility Compliance →Enter your website and get a free 60-second performance, SEO & accessibility report.
~60 seconds · No login