Reject Buttons That Move, Shrink, or Fade: Reading a Banner Like a Regulator
Some banners make rejection physically harder: the button is smaller, lower contrast, in a second layer, or it moves. How those patterns are assessed, and a plain checklist for spotting them on your own site.
Open your own banner on a phone and try to reject. Count the taps, and compare them with the one tap that accepts. Then look at the two buttons as a designer would: size, colour, contrast, position, wording. If accepting is a large solid button saying accept all and rejecting is a pale link saying manage options, the banner is steering. European regulators have a name for this family of tricks, deceptive design patterns, and the EDPB published guidelines on them that describe interface patterns in detail rather than in the abstract.
None of this requires intent. Most banners that steer were not designed by anyone trying to break the rules. They are templates, bought or free, built by a vendor whose commercial pitch includes high opt in rates. A vendor boasting about consent rates is telling you something about how their banner achieves them, and the liability for the interface sits with the store that ships it.
The patterns that keep appearing
Asymmetry first: more steps, smaller target, or a settings layer to reach reject, against one click to accept. Then visual interference: colour and size used to make one option dominate. Then wording: confirm shaming, such as no thanks, I hate saving money, and double negatives in toggle labels that make the off state hard to identify. Then obstruction: a close control that appears after a delay, a banner that reopens when a page changes, or a settings panel where the save button sits below a long list while accept all is pinned at the top.
Each pattern alone can look trivial. Regulators tend to assess the combined effect on an ordinary user, on a small screen, in a hurry, which is the realistic condition. A pattern that a lawyer can navigate in a desktop review can still be a barrier for the shopper on a train with one thumb free. Test in that condition, with the device and the time pressure, not in a design review.
Toggles, defaults, and the settings layer
Inside a preference centre, every non essential category should start off, and the toggles should say what state they are in without relying on colour alone. Pre enabled marketing and analytics toggles inside a settings layer are one of the most frequently criticised patterns in enforcement and in guidance, because the shopper who made the effort to open settings is then opted in by default unless they notice and reverse each one.
Labels matter here too. Legitimate interests used as a parallel basis for the same processing, with another set of toggles most people do not understand, is a pattern regulators have pushed back on in the advertising framework context. If your settings layer needs a paragraph of explanation to operate, simplify the choices rather than lengthening the explanation.
A ten minute self audit
Screenshot the banner on mobile and desktop. Count clicks to reject all and to accept all; they should be equal. Check that reject is on the first layer, in a button of comparable size and contrast. Open settings and confirm every non essential toggle starts off. Reject, then browse the site and confirm the banner stays away and non essential tags stay off. Return through the footer cookie link and confirm you can change the choice as easily as you made it.
Write the findings down with dates, including the faults. If you later change vendor or template, that record shows the store checked and acted. If you find a pattern you cannot fix in your current tool, that is a concrete requirement to take into the next procurement decision, instead of a vague feeling that the banner is a bit pushy.
Related reading: Consent for Embedded Video and Marketing Pixels, Consent Record Retention: How Long to Keep Proof and Why, Consent for Embedded Reviews, Social Widgets, and Comment Tools.
Frequently asked questions
Is it illegal to put reject in a settings layer?
Regulators in the UK and EU have said that making rejection harder than acceptance undermines freely given consent, and a reject option only available in a second layer is the classic example. The French regulator has fined large companies over exactly this asymmetry. Put reject all on the first layer.
What if my consent platform cannot be changed?
Most platforms offer a compliant template or settings for button order, colour, and first layer rejection. Check the configuration before assuming you need a new vendor. If the tool genuinely cannot present an equal reject option, that is a documented reason to change tools.
Do high consent rates mean my banner is non compliant?
Not by themselves. Plenty of audiences accept willingly. But a very high rate combined with an asymmetric design invites the question of whether the design produced the rate. If rejection takes more effort than acceptance, the rate is not good evidence of genuine preference.
How often should consent choices be renewed?
There is no single statutory period for cookie consent in the UK or EU, but guidance commonly points to reviewing after a set period, often around six to twelve months, and whenever purposes or vendors change. Whatever period you choose, apply it consistently and record it in your cookie information.