← All cases
Case 003

Cookie banners: click the same box on every site, forever

Shut up and give me the cookie. Whether you want to accept or reject, you're forced to decide again on every site you've ever visited — a choice a browser could make once.

3000 9000
human lifetimes wasted per year
best single estimate: ~4800
Difficulty to fix moderate
⚠ Source & extrapolation to firm up
WebGovernment

Facts cited

Time EU spends clicking cookie banners 575 million hr / year Legiscope, 2023 (widely cited)
…in full-time-worker terms ≈ 275,000 FTEs same analysis
Websites showing a banner ~85% avg user sees ~1,020 / year
Cookie banners that FAIL GDPR anyway 80–90% compliance studies
Internet users worldwide ~5.5 billion ITU, 2026

Assumptions low → high

What we can't measurelowlikelyhigh
EU: hours per user per year (this one is measured, not guessed)The 575M-hour figure ÷ ~410M EU internet users. Shown as a band for the seconds-per-banner uncertainty. 1.0 hr 1.4 hr 2.2 hr
Rest of world: hours per user per yearBanners are near-universal, but enforcement is lighter outside the EU. Deliberately kept below the EU rate. 0.3 hr 0.5 hr 1.0 hr

Who can fix it

Owner
Browser makers · W3C (Global Privacy Control) · EU (ePrivacy Regulation)
Channel
globalprivacycontrol.org · browser issue trackers · EU ePrivacy consultation

The web’s most-clicked button isn’t Buy, or Like. It’s Accept. Every site, every device, every time your cookies clear — the same box, the same shrug, the same ten seconds gone. And here’s the kicker: 80–90% of these banners don’t even comply with the law they exist for. Maximum friction, near-zero privacy. The worst of both worlds.

The math

The EU part isn’t even an estimate on my side — it’s a measured, cited figure:

575M hours / year (EU) = ~900 lifetimes / year — just the EU

Extended worldwide at a deliberately conservative rate (the rest of the world kept below the EU’s per-capita clicking), the napkin lands at:

EU 575M hr + 5.1B users × 0.5 hr = ~4,800 lifetimes / year

Even the low end — ~3,000 lifetimes a year — is fifty times cases 001 and 002 combined. This is the monster. Cookie banners are civilisation’s single largest self-inflicted UI paper-cut.

Two things to firm up before this number is quoted as settled.

  1. The 575M-hour figure is widely repeated but traces back to a single analysis (Legiscope), not a peer-reviewed study. Its own assumptions (banners per user, seconds each) need reading before we lean on it — repetition across blogs is not independent confirmation.
  2. The worldwide extension is our extrapolation, not anyone’s measurement. The EU portion (~900 lifetimes/yr) is the only part with a cited source behind it.

Per the methodology, the honest floor here is the EU figure. The global number is a labelled estimate, and it is doing the heavy lifting in the headline.

The polite banner sits at the edge of the screen and lets you read. Its nastier cousin greys out the entire page and blocks it until you engage. This crosses a line: the content is already loaded — it exists behind the dimming — and it’s being deliberately withheld until you comply. That’s not a request, it’s hostage-taking. “Please consent” becomes “consent or you get nothing.”

It almost always ships with the reject-parity violation: Accept all is one bright button, while Reject is buried behind “Manage preferences → toggle everything off → confirm.” GDPR explicitly requires reject to be as easy as accept — so the wall is usually illegal as well as gross, and doubly pointless, because no one cornered into clicking has meaningfully consented to anything.

Same through-line as the rest of this site: the system refuses to let no be easy, and makes the adverse action the path of least resistance.

The banner pretends to be a choice. In practice it’s a reflex — everyone clicks the biggest button to make it go away, which is exactly why 80–90% of implementations quietly violate GDPR’s “reject must be as easy as accept” rule. Nobody is meaningfully consenting. The time is pure deadweight loss: no privacy delivered, hours destroyed.

The honest reframing: consent shouldn’t be a per-site interruption — it should be a setting you choose once. That serves everyone:

  • People who don’t care (accept-all) never see a banner again.
  • People who do care (reject-all) get it enforced everywhere, automatically — stronger privacy than clicking “reject” on 1,020 sites a year and missing some.

Either way, nobody clicks. The friction, not the privacy, is the enemy.

killthecookiebanner.eu — a coalition of seven civil-society organisations including noyb, EDRi, the EFF and BEUC — is campaigning for exactly the fix described below: browser-level automated signals that carry your privacy preference, so you set it once instead of clicking a box on every site. They’re backing the European Commission’s autumn 2025 proposal and asking people to contact their MEP or national representative.

Their sharpest number, and it demolishes the consent fiction: up to 90% of users click “yes” — while only about 3% actually want tracking. (Figures as cited by the campaign; we haven’t traced them to the underlying study.)

Worth repeating their own caveat, because it’s careful: they explicitly do not endorse the rest of the Digital Omnibus reform package, which they consider likely to weaken people’s rights. Supporting the browser-signal fix isn’t the same as supporting everything it’s bundled with.

The fix already exists — it just isn’t mandated

  • Global Privacy Control (GPC) is a browser-level “here’s my standing preference” signal. California’s CCPA legally requires businesses to honour it. It works. It’s just not mandatory in the EU, so sites ignore it and show the banner anyway.
  • The dead Do Not Track header was this idea a decade early; industry killed it by choosing not to honour it. The lesson: the signal needs legal teeth, not goodwill.
  • The EU’s own ePrivacy Regulation has been trying to move consent to the browser for years.

So the fix is difficulty: moderate — not one line of code, but not moonshot either. It’s a browser default plus a regulation that says “honour the browser signal, banner forbidden.” The technology is shipping today.

Second-order waste — directional, only pushes up

  1. Consent fatigue bleeds into everything. Trained to reflexively click the big button, users click through real security and permission prompts the same way.
  2. The banner-industrial complex. An entire consent-management SaaS industry now bills businesses to generate the friction. Money and engineering spent manufacturing the paper-cut.
  3. Page-load and layout-shift tax. Consent scripts are among the heaviest third-party blockers on the web — slower pages for 5.5 billion people, on top of the click.

Difficulty to fix: moderate

The web’s most-clicked button is “Accept.” Not because anyone consents — because it’s the fastest way to make the box disappear. Move the choice into the browser, make it once, and hand ~4,800 lifetimes a year back to humanity. Just give me the cookie.