In 2026, cookie banners that offer a truly equal Accept and Reject button still lose 60% of their consent-gated visit data, according to etracker’s 2025 Consent Benchmark Report of 500 cross-industry websites. That is the modern, traffic-based answer. The older, click-by-click answer is more precise and just as revealing: in a 2019 field study that logged 82,890 real visits, only 13% to 14% of people actively clicked Decline, even when Decline sat right next to Accept.
What percentage of visitors reject cookies when given a fair choice?
On websites that offer a genuinely equal Accept and Reject button, the answer is close to a coin flip that tips toward reject. etracker’s 2025 Consent Benchmark Report, which analyzed 500 cross-industry websites in June 2025, found a median consent rate of just 40% among sites with a top-level reject button sized, colored, and positioned the same as Accept. That means a median 60% of visitors do not consent, and the effect is not subtle: more than one in five compliant websites see their consent rate fall below 10%.
The report frames this as data loss rather than rejection outright, since the 60% figure lumps together people who click Reject and people who never interact with the banner at all. For the flip side of this number, the accept-rate trends and banner-design breakdown live in Cookie Consent Statistics 2026. The next section separates the click from the silence using data logged from real visits, not aggregated traffic.
A fair choice does not produce a fair split. Most visitors end up on the no-consent side of the ledger.
How many people actually click Reject versus never respond at all?
Traffic-based consent rates blend two very different behaviors: an active click and simple inattention. Utz et al.’s 2019 field study, which logged 82,890 real visits to a German ecommerce site over four months, is still the largest dataset that separates them. In the binary condition, where Accept and Decline sat side by side with equal styling, only 14.3% of mobile visitors and 13.2% of desktop visitors actively clicked Decline.
Figure 1: What mobile visitors actually did when Accept and Decline had equal styling. Source: Utz, Degeling, Fahl, Schaub, Holz, ACM CCS 2019.
Desktop visitors behaved differently. Accept drew just 20.9% of desktop clicks, and 65.9% of desktop visitors did nothing at all, nearly twice the mobile no-action rate. Either way, an explicit reject click is a minority behavior across both device types. Most of what gets counted as rejection in aggregate traffic data is really silence, not a decision.
Clicking Reject is rare. Not clicking anything at all is the norm.
Does the design of the reject button change how many people click it?
Yes, and the effect shows up immediately. When Utz et al. highlighted the Accept button in a contrasting color while leaving Decline as a plain unstyled link, the mobile decline rate dropped from 14.3% to 10.1% and the desktop decline rate dropped from 13.2% to 11.9%. Accept clicks rose to 45.5% on mobile in the same condition, while desktop Accept clicks stayed almost flat at 20.1%, since desktop visitors mostly stayed inactive either way.
Figure 2: A highlighted Accept button lowers the decline rate on both device types. Source: Utz et al., ACM CSS 2019.
This single design change, no new legal text and no new consent categories, just a color and a font weight, moved several percentage points of decisions. Regulators have taken notice. etracker’s 2025 audit of 500 live websites found that 48% still nudge visitors toward Accept through button color, contrast, or by rendering Reject as a plain text link instead of a matching button, years after guidance first told them not to.
The reject rate is not a fixed fact about users. It is a setting a site owner controls.
Are more websites offering a genuine reject option over time?
Slowly, yes. etracker’s longitudinal comparison of the same website sample found that the share of sites offering a reject option at the top level, meaning no extra click into a settings menu required, rose from 80% in November 2023 to 87% in May 2025. The share offering a genuinely equal-prominence Accept/Reject button, the design regulators actually require, nearly doubled across the same period: 27% to 52%.
Figure 3: Reject-button prevalence and design fairness, 2019 to 2025. Sources: Utz et al. 2019, etracker 2023 and 2025.
Progress has a ceiling, though. Even in the most recent data, only 15% of the top 10,000 EU websites run a banner that clears the full compliance bar, meaning equal prominence, granular purpose controls, and a working reject path, according to CHI 2025 research analyzing 254,148 sites. A rising share of sites with a fair button does not yet mean a majority of sites are compliant. For the full count of how many sites show a banner at all, see how many websites have a cookie banner.
The direction is right. The finish line is still far off.
What makes a Reject option legally fair, not just present?
German supervisory authorities have published a specific, testable standard rather than a vague principle. Their joint guidance for digital service providers states that the option to withhold consent must be presented as a clearly equal alternative to accepting, matched in size, color, contrast, and typography, whenever a same-effort Accept button sits next to it. A German administrative court applied that standard directly: the Verwaltungsgericht Hannover ruled in March 2025 that a reject option must sit on the first layer of a banner, not behind a settings click.
Figure 4: The two-question test German regulators apply to a reject option. Source: German joint DPA guidance, cited in etracker’s 2025 report.
That test explains why etracker’s compliant-banner group and non-compliant group post such different consent numbers. Sites that pass the test see a median 40% consent rate. Sites that do not average 14 percentage points higher, roughly the gap you would expect if part of that extra consent comes from people who could not find a real way to say no.
A visible Reject button and a fair Reject button are not the same thing, and regulators are now testing for the difference.
If there is no reject button, do visitors accept instead, or just leave?
Mostly, they leave. In Utz et al.’s no-option condition, where the banner offered only an X to dismiss and no way to actively decline, 52.5% of mobile visitors closed it without engaging further, compared with only 14.3% of mobile visitors who clicked Decline in the equal-button condition.
Figure 5: Removing the reject option raises silent dismissals, not consent. Source: Utz et al., ACM CCS 2019.
Removing the reject button does not convert reluctant visitors into willing ones. It mostly converts an active decline into a passive dismissal, which is worse for a site owner, since a dismissal with no recorded choice is the weakest possible basis for any tracking that follows. The safer path is to offer a real reject option and pair it with a document that matches what the banner actually does. You can generate a compliant cookie policy that states the reject behavior plainly instead of leaving it implied.
No reject button does not mean more consent. It means less data about what people actually wanted.
Reject rates at a glance
The clearest way to see the design effect is side by side.
| Metric (etracker Consent Benchmark 2025, 500 sites) | November 2023 | May 2025 |
|---|---|---|
| Sites with a reject option at the top level | 80% | 87% |
| Sites with a genuinely equal Accept/Reject button | 27% | 52% |
| Sites still nudging visitors toward Accept | not disclosed in this wave | 48% |
| Average visit-data loss under compliant design | not measured in this wave | 60% |
Source: etracker Consent Benchmark Report, 2023 and 2025 waves.
The Bottom Line
The honest answer to what percentage of users reject all cookies depends on what counts as rejection. Counted as an explicit click, it is small: 13% to 14% in the largest field study that has logged it. Counted as the share of visitors who end up not consenting once ignored banners and closed notices are added in, it is large: a median 60% on sites that offer a genuinely fair choice. Both numbers point to the same conclusion for a site owner. Most visitors will not actively opt in once the thumb comes off the scale, and removing the reject button entirely just trades an honest decline for a silent one. A cookie policy that plainly states what happens on accept and on reject, matched to a banner that actually offers both with equal weight, is the version of this that survives a regulator’s second look.
Frequently Asked Questions
What percentage of users click reject on cookie banners? In a large-scale field study logging 82,890 real visits, only 13.2% to 14.3% of visitors actively clicked Decline on an equally prominent Accept/Reject banner. Most of the rest never interacted at all (Utz et al., ACM CCS 2019).
Do most people reject all cookies when given a fair choice? Not through an explicit click, but most do not end up consenting. On websites with a genuinely equal Accept/Reject button, the median consent rate is just 40%, meaning a median 60% do not consent, according to etracker’s 2025 Consent Benchmark Report of 500 websites.
Does the design of the reject button change how many people click it? Yes. Highlighting the Accept button cut the desktop decline rate from 13.2% to 11.9% and the mobile decline rate from 14.3% to 10.1% in Utz et al.’s 2019 field data, and 48% of websites still nudge visitors toward Accept in 2025 (etracker).
Are more websites offering a genuine reject option over time? Yes. The share of websites with an equally prominent Accept/Reject button rose from 27% in November 2023 to 52% in May 2025, according to etracker’s Consent Benchmark Report.
Sources and References
- etracker GmbH. (2025). “Consent Benchmark Report 2025.” 500 cross-industry websites, data through June 2025.
- Utz, Degeling, Fahl, Schaub, Holz. (2019). “(Un)informed Consent: Studying GDPR Consent Notices in the Field.” ACM CCS 2019. Field experiment, 82,890 unique visitors, November 2018 to March 2019.
- Nouwens, Kristensen, et al.. (2025). “A Cross-Country Analysis of GDPR Cookie Banners and Flexible Methods for Scraping Them.” CHI 2025, ACM. 254,148 websites across 31 countries.
- Advance Metrics. (2024). “Cookie Behaviour Study, 5 Years After GDPR.” Observed first-click behavior of 1.2M+ visitors, data period 2023.
Note: All figures verified as of August 2026. Consent behavior and banner design shift quickly, so headline figures are refreshed at least twice a year.
