Your cookie consent rate is the share of visitors who answer your banner by accepting at least one optional category, such as analytics or marketing. Published research puts it well below 100%. In a representative Bitkom survey of 1,013 internet users in Germany, published on 25 March 2024, 45% said they accept all cookies as a rule, 21% said they reject everything they can and 33% said they choose category by category. Banner design moves the figure most: in a 2020 experiment by Nouwens and colleagues, removing "Reject all" from the first layer raised consent by 22 to 23 percentage points.
Benchmark tables that circulate online rarely name their data. The figures below come from named studies, each with its sample and date, so you can judge how far they apply to your own site.
A low consent rate leaves gaps in your data
Every visitor who declines cookies creates a gap in your analytics data. That gap compounds across sessions, channels and campaigns until your reporting tells a story that barely resembles reality.
The size of that gap depends on your audience, your traffic and how your banner is designed. Knowing which of these you can change, and which you cannot, is the first step toward realistic data-collection expectations and better decisions about your analytics.
Article 5(3) of the ePrivacy Directive requires prior consent before placing non-essential cookies on a visitor's device. Under the GDPR, that consent must be freely given, specific, informed and unambiguous. These rules shape consent rates everywhere the ePrivacy Directive applies.
What the published studies measured
Three studies give consent figures with a stated method, sample and date. They measured different things, so read each row on its own terms rather than averaging them.
| Study | Where and when | Sample | What it found |
|---|---|---|---|
| Bitkom Research survey, published 25 March 2024 | Germany, January 2024 | 1,013 internet users aged 16 and over (self-reported) | 45% accept all cookies as a rule, 21% reject all, 33% choose. Of those who choose, 10% allow advertising cookies |
| Utz et al., ACM CCS 2019 | A German-language e-commerce site, November 2018 to March 2019 | 82,890 visitors (observed) | With every category ticked in advance, 83.55% of visitors who chose accepted. With none ticked, 0.16% did |
| Nouwens et al., ACM CHI 2020 | 680 of the top 10,000 UK websites (scraped September 2019), plus a field experiment | 40 participants, 1,280 answers (observed) | 12.6% of sites showed "Accept all" and "Reject all" on the same page. Participants chose "Accept all" 55.2% of the time and "Reject all" 34.1% |
The gap between the first two rows is the point. People say one thing in a survey and click another on a real banner, and pre-ticked boxes inflate acceptance to a level no longer allowed: since the Court of Justice's Planet49 judgment of 1 October 2019 (case C-673/17), a pre-ticked box is not valid consent.
Industry benchmark tables rarely show their data
None of the studies above breaks consent rates down by industry, and no peer-reviewed dataset found for this article does. Tables that put healthcare at 60% or ad-supported media at 25% usually come from a vendor's own customer base, without the sample, the countries or the banner designs behind each figure.
The reasons a sector might sit higher or lower are still worth knowing. Healthcare and financial sites can offer a clear value exchange or the trust of a regulated brand. Ad-supported media carries heavy third-party tracking and a long history of intrusive advertising. On e-commerce sites, returning customers who are signed in tend to accept more readily than first-time browsers, so your rate shifts with the mix of new and returning traffic. Treat these as hypotheses to test in your own data, not as targets.
The benchmark that matters most is your own rate before and after each change.
Germany and the EU: what the data shows
Germany is the best-studied market. The Bitkom survey found that 76% of internet users in Germany are annoyed by cookie banners and 68% say they would rather not deal with them at all, a clear sign of consent fatigue. 58% said they delete cookies in their browser regularly. Among the third who choose category by category, 70% allow functional cookies such as language settings, 47% allow cookies for improving the product and only 10% allow advertising cookies.
The observed data from Utz and colleagues points the same way. On their German e-commerce site, a confirmation-only notice with a highlighted "Accept" button was accepted by 50.8% of mobile and 26.9% of desktop visitors. When "Accept" was shown as a plain text link instead, those figures fell to 39.2% and 21.1%.
For the United States the picture is different in kind. Most state privacy laws, including the California Consumer Privacy Act, use an opt-out model, so many sites serving mainly US visitors show no opt-in banner at all and have no consent rate to measure in the EU sense.
If your website serves several regions, your overall rate blends these models. Segmenting consent data by country gives a clearer picture: the Overview tab of the Analytics page in Kukie.io shows the accept, reject and custom rates next to your top countries.
Desktop and mobile visitors behave differently
Device type changes how often visitors answer at all. In the first experiment by Utz and colleagues, only 5.5% of visitors with screens of 1,920 by 1,080 pixels or larger interacted with the notice, against 25.6% on smaller screens. On a large desktop screen a small banner is easy to ignore. On a phone it covers the content.
Mobile screens force banners into smaller spaces, which makes it tempting to hide the reject option behind an extra tap. Regulators, including the CNIL, have fined sites where refusing took more steps than accepting. A banner that covers most of a phone screen and makes visitors scroll to find "Reject all" invites both regulatory scrutiny and frustration. Progressive disclosure, where the first screen presents the main choices and detail sits behind a secondary tap, works well on small screens without resorting to dark patterns.
Traffic source has a measurable impact
Your consent rate is not uniform across channels. No large published study has measured the effect of each traffic source, so the table below lists tendencies to check in your own data rather than measured averages.
| Traffic source | Tendency to check | Likely reason |
|---|---|---|
| Google organic search | Higher acceptance | Intent-driven, focused on content |
| Social media referrals | Moderate acceptance | Casual browsing, lower commitment |
| Direct traffic | Higher acceptance | Returning visitors, brand familiarity |
| DuckDuckGo organic | Lower acceptance | Privacy-conscious user base |
| Email campaigns | Higher acceptance | Existing relationship with sender |
| Paid advertising | Moderate acceptance | Mixed intent, first-time visitors |
Someone who actively chooses a privacy-focused search engine is more likely to decline tracking on your site. Direct traffic and email referrals tend to arrive with a relationship to your brand already in place. For a deeper look at how referral sources shape consent behaviour, see traffic source and consent rates.
Banner design is the variable you control
Audience, geography and traffic source are largely outside your control. Banner design is not, and it produces the largest measured effects.
In the field experiment by Nouwens and colleagues, published at the ACM CHI conference in April 2020, removing the "Reject all" button from the first layer raised the probability of consent by 22.0 percentage points for a full-screen notice and 23.1 points for a banner. Showing more granular choices on the first layer lowered it by 8.8 to 20.0 points, depending on whether purposes, vendors or both were listed. The same study found that 93.1% of interactions stayed on the first layer: anything behind a "More options" link was rarely seen.
How first-layer design changed consent in one experiment
Pre-selection has an even larger effect, which is why it is no longer allowed. Utz and colleagues saw acceptance of a category notice rise from 0.16% with every box unticked to 83.55% with every box ticked in advance. Position matters too: on their site, a notice in the bottom-left corner drew the most interactions (37.1% of visitors), while small bars at the top or bottom of the screen drew 2.9% and 9.6%.
A banner with "Reject all" next to "Accept all" will usually record a lower acceptance rate than one that hides the refusal. That gap is what valid consent costs, and the experiments above suggest it can reach 20 points or more. Clear category descriptions, plain-language copy and a position visitors notice are the levers that remain.
What low consent rates mean for your analytics
When half your visitors decline cookies, your analytics data becomes a partial picture. Google Analytics 4 shows only consented data in its reports unless you use blended reporting, so traffic, conversion and attribution figures all carry a consent-shaped blind spot.
Google's Consent Mode v2 tries to fill this gap through behavioural modelling: it estimates the actions of visitors who declined, based on patterns from those who accepted. Google only models a property that meets its thresholds, at least 1,000 events a day with analytics_storage denied for at least 7 days, and at least 1,000 users a day with it granted on at least 7 of the previous 28 days. Modelled data helps, but it is an estimate, not a measurement, and the fewer consented visitors you have, the less there is to model from.
Sites with low consent rates can supplement their analytics with privacy-preserving analytics tools that work without cookies. Tools like Plausible, or Matomo in its cookieless configuration, can give baseline traffic data for all visitors, regardless of consent status. Check what your data protection authority says about consent-free measurement before you rely on it.
How to improve your consent rate without dark patterns
The goal is not to trick visitors into accepting but to remove unnecessary friction and explain the value clearly.
Start by auditing your current banner. Run a cookie scan to see which cookies your site sets, then make sure your banner describes each category accurately. Visitors are more likely to consent when they understand what they are agreeing to. Vague descriptions like "improving your experience" tell visitors nothing, while specific ones like "remembering your language and basket contents" let them decide.
Timing matters. A banner that appears after a brief delay or on scroll gives visitors a moment to see whether the site offers what they came for, as long as every non-essential script stays blocked until they choose. Test different approaches within legal bounds, as outlined in the guide to improving cookie consent rates.
Reducing the number of non-essential cookies on your site also helps. Fewer cookies mean a simpler choice and less reason for visitors to decline.
Short answers on consent rates
What is a good cookie consent rate for my website?
There is no published standard. In Bitkom's 2024 survey of internet users in Germany, 45% said they accept all cookies as a rule and 21% reject all. Compare your own rate before and after each banner change rather than against an industry table.
How do I calculate my consent rate?
Divide the number of visitors who accepted at least one optional category by the number of visitors who saw the banner. Some tools divide by the visitors who answered instead, which gives a higher figure, so check which base a benchmark uses before you compare.
Why are healthcare websites said to have higher consent rates?
No published study measures consent rates by industry, so the higher figures quoted for healthcare come from vendor data. A plausible reason is a clear value exchange: cookies for appointment reminders or patient portals are easier to justify to visitors.
Do mobile users accept cookies more than desktop users?
In the study by Utz and colleagues on a German e-commerce site, they did: a notice with a highlighted Accept button was accepted by 50.8% of mobile visitors and 26.9% of desktop visitors. Desktop visitors on large screens often ignored the notice altogether.
Can I improve my consent rate without using dark patterns?
Yes. Clear category descriptions, plain-language copy, sensible timing and fewer unnecessary cookies all help. Hiding the reject option raised consent by 22 to 23 percentage points in the 2020 experiment by Nouwens and colleagues, which shows how much of a high rate can come from design that does not give a free choice.
What happens to my analytics when visitors reject cookies?
Visitors who reject cookies are not tracked by tools like Google Analytics 4, which reports only consented data unless blended reporting is used. Consent Mode v2 can model their behaviour once a property meets Google's thresholds, but a low consent rate leaves less data to model from.
See how your own visitors answer the banner
Published benchmarks give you a range. Your own figures tell you whether a change worked. Kukie.io records every choice in the Consent Log and shows the accept, reject and custom rates on the Analytics page of each site, so you can compare your rate before and after each change to your banner.
