In one paragraph
Reject button less prominent than accept means the observed consent controls differ on at least one dimension the scanner measures. Review the evidence and the complete choice design rather than treating every difference as unlawful or every equal-size pair as fair. European authority guidance requires context-specific assessment. A filled versus outlined control is not, by itself, a complete legal or accessibility conclusion.
Why it matters
Consent design can steer users through wording, colour, contrast, position, visibility, and effort. A technically present refusal that people cannot reasonably find or read can undermine meaningful choice.
The EDPB taskforce expressly rejected imposing one general colour or contrast standard for all banners and called for case-by-case review, while identifying clearly misleading examples such as virtually unreadable alternative text. It did not establish the scanner's ninety-percent area threshold as a legal minimum.
Visual parity is also different from functionality. A refusal can look equally prominent yet fail to apply the choice, while minor typography differences may have little actual significance. Test both the visible choice and the resulting storage or requests.
How WebAuditOS evaluates this finding
The supplied mobile-snapshot rule compares the first recognised acceptance and refusal actions. It warns at low severity for its documented smaller-area threshold, transparent-versus-filled treatment, link styling, smaller font, or lighter weight.
No measured difference passes this comparison; no eligible pair or unreadable evidence can be ungraded under its conditions. It does not measure every text contrast, border, position, scroll requirement, language meaning, desktop state, or actual rejection. Its strict font comparisons can flag small differences that require judgment.
How to fix it
- Reproduce the first-layer choice in a fresh relevant context. Record the recognised actions and measurements, and confirm the naming heuristic has selected the intended accept and refuse controls.
- Review the whole layout, including text and background contrast, borders, focus, position, visibility without scrolling, mobile and desktop, and translated wording. Do not rely only on bounding-box area.
- As a practical baseline, make genuine alternatives clearly understandable and comparably usable. Similar styles can reduce ambiguity, but do not present identical pixels as a universally mandated design.
- Fix misleading text, hidden alternatives, or effectively unreadable controls. A technically valid link in a long paragraph may not provide an adequately noticeable refusal path.
- Test keyboard order, touch targets, zoom, and supported assistive output. A visual repair that leaves refusal unnamed or inaccessible is incomplete.
- Activate refusal and verify actual behaviour before choice, after choice, on navigation, and after withdrawal. Style equality does not authorise uncontrolled collection.
- Review local authority requirements and the specific facts with appropriate expertise where needed. The taskforce report is a coordinated interpretive baseline, not a stand-alone greenlight for every banner.
- Rescan and keep relevant measurements as evidence of the observed state. Document manual conclusions rather than calling the warning a definitive legal infringement or the pass proof of consent validity.
What the fix can break
Changing styles can affect contrast, focus visibility, wrapping, and translated labels. Test complete states rather than a single desktop mock-up.
An outline can be deliberate and readable, while an equal-area button can still be visually buried. Avoid reducing design judgment to one metric.
Do not solve a warning by hiding the reject control from the detector or changing its name to an unrecognised word. Improve the real choice.
Questions
Is ninety percent an EDPB legal threshold?
No. It is the supplied scanner's area comparison rule, not a statutory design formula.
Does a different background automatically invalidate consent?
No automatic verdict follows from that fact alone. Assess the full presentation and actual context.
Does a pass prove refusal works?
No. This rule compares selected visual measures without completing the rejection flow.
Which related checks help?
Review direct refusal, preselected choices, and the cookie consent guide.