Accessibility is about real people completing real tasks. Automated inspection can spot missing labels and structural clues, but it cannot experience a website. Millwards Online therefore reports a narrow category score and states clearly that further human testing is needed.
What this accessibility check is for
Small structural omissions can create immediate friction. A form field without an associated label may be unclear to a screen-reader user. An unnamed icon button can announce no useful action. A missing page language can make pronunciation less reliable. A missing mobile viewport can force awkward scaling on smaller screens.
These are useful signals to check because they can often be detected consistently in returned HTML. They are also only a small part of accessibility. The audit is designed to identify obvious foundations and make its evidence visible—not to issue a pass certificate or legal opinion.
Use it when a website is launched or updated, when a new form or navigation is introduced, or when you need a starting checklist before commissioning a broader review. The same accessibility category is included within every free small-business website audit.
Accessibility signals included in the scan
- Mobile viewportLooks for a viewport declaration that supports responsive presentation on mobile devices.
- Document languageChecks whether the root HTML element declares a detectable language such as
en-GB. - Form labelsEstimates whether detectable inputs are connected to labels or have an appropriate accessible name.
- Button namesChecks whether buttons expose text or a detectable accessible label rather than an unnamed control.
- Keyboard skip linkLooks for a link that lets keyboard users bypass repeated navigation and reach main content.
- Image alternativesA separate content rule measures the proportion of images with detectable alternative text attributes.
These checks assess presence, not full quality. An image can have an alt attribute that is unhelpful. A label can exist but use confusing language. A skip link can be present yet hidden incorrectly. A button name can be technically available while the action remains ambiguous. The report provides evidence to investigate, not a final human judgement.
How form-label coverage is interpreted
The parser looks for conventional label relationships and accessible naming patterns in the returned document. A full ratio receives a stronger result; partial coverage can produce a warning. Forms inserted after JavaScript runs or displayed only after interaction may not be available to this server-side check.
Why image alternatives sit in the content category
Alternative text depends on the purpose of the image and surrounding content, so the current model includes its detectable ratio among content signals. It still matters to accessibility, and the complete report can make a missing attribute visible. Decorative images may correctly use an empty alternative; automation cannot always understand that context.
What users receive and how the score works
The free result shows the accessibility category score and public detail for the mobile viewport check. Other accessibility findings are counted but remain locked. The complete £29 report reveals each applicable result with its detected evidence, likely impact, recommendation and severity.
Every rule earns a fraction from zero to one and has a weighting. The accessibility category is the weighted average of its five rules; image alternatives contribute through the content category. The overall score combines all categories. A pass means the relevant pattern was detected strongly enough for that rule—not that the feature was tested successfully with assistive technology.
The example report uses fictional data to show how an unlabelled field could become a prioritised finding. The methodology publishes the general score bands and limitations.
Why this cannot be a complete accessibility audit
No automated tool can prove that every person can use a website. Millwards Online does not claim WCAG conformance, legal compliance or certification.
The scanner does not operate the site as a person using a keyboard, switch, voice input, screen reader, magnification or other assistive technology. It does not judge whether focus order is logical, focus is visible, error messages are helpful, a modal traps focus correctly or a multi-step journey can be completed.
It also does not fully assess colour contrast across every state, text resizing and reflow, captions and transcripts, flashing content, motion controls, reading level, link purpose, target size, cognitive load or the accessibility of PDFs and embedded third-party tools. JavaScript-only interfaces and content behind consent or interaction may not be present in the HTML it receives.
Legal duties depend on context and jurisdiction. A detected privacy link or a high accessibility score is not legal advice. If compliance or a high-impact user journey matters, arrange a suitable expert review and testing with disabled users.
What to do after the automated scan
- Verify the evidence. Find the named field, button or document signal on the live page and confirm the report saw the intended version.
- Fix clear structural gaps. Connect visible labels, name controls, declare the page language and provide a working skip link where appropriate.
- Test the whole task by keyboard. Move forward and backwards, activate controls, recover from errors and confirm focus remains visible.
- Use screen-reader and zoom checks. Confirm headings, landmarks, control names, order, announcements and reflow make sense.
- Review content and visuals. Test contrast, alternatives, captions, instructions and plain-language clarity in context.
- Include people, not only tools. For important services, user testing and an appropriately scoped professional audit add evidence automation cannot supply.
After changes, a fresh scan can confirm that the detectable structure changed. Monthly monitoring can compare future automated results, but it does not replace recurring human accessibility review.
Website accessibility audit questions
Is this a complete WCAG accessibility audit?
No. It checks a small set of detectable structural signals in the submitted page. It cannot establish WCAG conformance or replace keyboard, screen-reader, zoom, contrast, cognitive and user testing.
Can an automated accessibility score prove that a website is accessible?
No. A good category score only means the page covered the defined automated rules well. Many important barriers require context and human judgement.
Will the audit change my website?
No. It reads publicly returned HTML and does not log in, submit forms or alter the site.