Cookie banner requirements:
what we measure on your banner

In 2023 the European data protection authorities jointly recorded which banner designs they check as a minimum threshold. The scan opens your banner like a visitor, measures each of these points and names the source and the setting in your banner tool for every finding.

Check your website for free

Your account after the first check

Account view of a sample shop: points to fix, notes, services, history

Sample shop. Every point opens its fix with the steps for your tool.

Free for 14 days, no credit card. Results usually in under two minutes.

What we measure on the banner

Each point is a measurement on the opened banner, with its place in the report of the Cookie Banner Taskforce of the European Data Protection Board (EDPB) of 17 January 2023. Whether a banner is sufficient in a given case is for the authority to decide; we show what we measured.

  1. 01

    Reject on the first layer

    Is there a reject option next to "Accept all", or only behind "Settings"? We read every button and link in the banner, including wording such as "Necessary only" or "reject here" in the text.

    EDPB Taskforce, paras 6 to 8

  2. 02

    Reject as a button, not a text link

    Does reject have its own surface or border like accept, or is it a link inside the text? We measure background, border and font of both.

    EDPB Taskforce, paras 11 to 14

  3. 03

    Reject inside the banner

    Is reject placed outside the banner, for example in the footer, while the accept buttons sit in the banner?

    EDPB Taskforce, para 14

  4. 04

    Contrast of the reject button

    We calculate the WCAG contrast ratio of text and background for reject and accept. Below 3:1 it is in the report, with the CSS line that copies the accept button's colours.

    EDPB Taskforce, paras 15 to 19; WCAG 2.1

  5. 05

    Pre-ticked boxes

    We open the settings and read every switch. Only strictly necessary cookies may be switched on.

    EDPB Taskforce, paras 9 and 10; CJEU C-673/17

  6. 06

    Legitimate interest in the settings

    Do purposes or vendors in the settings rely on legitimate interest, often with IAB TCF? We quote the sentence it appears in.

    EDPB Taskforce, paras 20 to 25

  7. 07

    A way back after choosing

    We accept in a fresh browser, then look on a second page, the cookie declaration and the privacy page for an icon or link that reopens the banner.

    EDPB Taskforce, paras 31 to 35; Art. 7(3) GDPR

  8. 08

    Continued browsing as consent

    Does the banner text say that continued use counts as consent, or that the website can only be used with consent? We quote the sentence word for word.

    EDPB Taskforce, para 13; EDPB Guidelines 05/2020, para 86

  9. 09

    Purpose named in the banner

    Does the text say what cookies are used for, such as statistics or advertising? A language model reads the text only when no purpose word is in it; without its confirmation there is no finding.

    EDPB Taskforce, paras 12 and 13

  10. 10

    Loading before the click and after "Reject"

    Alongside, the scan still measures which services load before consent and after "Reject", with a counter-check in a normal browser. More: <a href="/en/guides/cookies-before-consent">Cookies before consent</a>.

    Art. 5(3) Directive 2002/58/EC

Step by step

  1. Open the first layer

    The scan loads your home page in a fresh browser and looks for the banner: an accept button in a fixed frame that talks about cookies. Then it measures every button and link in it.

  2. Open the settings

    If there is "Settings" or "Customise", it clicks and reads every switch: name, on or off, locked or not.

  3. Accept, then move on

    In a second fresh browser it clicks "Accept all", opens a subpage and looks for the way back to the settings. If it finds none, it reads your cookie declaration and privacy page before anything goes in the report.

  4. A fix for your tool

    Every finding names what was measured, the source and the setting in your banner tool, and for Cookiebot, Usercentrics, Borlabs, CookieYes and OneTrust also the link code that reopens the banner.

  5. Every week again

    With a subscription the scan measures the banner again every week. If a banner tool update changes the buttons, the email says so.

Questions

Is a reject button required?

The EDPB Taskforce report records that the vast majority of authorities consider a banner without a reject option on any layer not compatible with valid consent (paras 6 to 8). Whether reject must be on the first layer is assessed differently by individual authorities. We measure where it is and show it yellow, never red.

Is this a legal opinion on my banner?

No. We measure facts on the opened banner and name the place where the authorities address the topic. Whether a banner is sufficient in a given case is for the competent authority to decide, which is why each of these findings is yellow in the report.

Where does the list come from?

From the report of the Cookie Banner Taskforce of the European Data Protection Board of 17 January 2023. It records what the authorities agreed on as a minimum threshold and is the common basis for banners across the EU. Plus the EDPB Guidelines 05/2020 on consent.

My banner tool says it is GDPR compliant. Why measure?

The tool provides templates and settings; how your banner looks is decided by the configuration on your website. We measure the banner your visitors see, and we sell no banner ourselves.

Is the check free?

Yes, without a credit card. Your website and email address start a 14-day trial: right away the points per area and the most important fix in your account, and once you confirm your address every finding with its location and the steps to fix it. The trial ends by itself.

What if the scan does not find my banner?

Some banner tools show no banner to automated checks. Then none of these findings is in the report, including none claiming something is missing; we do not get around any protection.

What a scan shows

It measures buttons, colours, switches, text and the way back; it does not judge whether your cookie list is complete or a cookie is correctly classed as necessary (paras 26 to 30), because that cannot be measured from outside. Whether your banner really blocks: Check your cookie banner yourself.

Check your website for free

As of 21 November 2026. We name the reference so you can read it yourself; this is not legal advice.

More guides

Check your website for free

Free, no email or credit card needed. The result is usually ready in under two minutes.

With email: every fix and a weekly check for 14 days.

  1. Result with the most important fix, no email needed
  2. With your email: every fix and a weekly check for 14 days
  3. The trial ends by itself, no credit card