VaferYes, a cookie banner costs you something — but a well-built one costs you single digits, a badly built one can cost you a quarter of your traffic, and most of what merchants panic about losing is reporting visibility, not actual sales. Here is what the published data really says about banners and conversion, why the numbers disagree so much, and the four levers that decide whether yours is cheap or expensive.
What the data actually says (and why it disagrees)
Published numbers are all over the place: average acceptance rates get quoted anywhere from around 30% to over 70%, bounce-rate impact from a few percent to 40%. The disagreement is mostly methodology: whether "acceptance" counts the visitors who ignored the banner entirely, whether the banner was a dismissible bar or a full-screen wall, and which country's visitors are being measured (US visitors accept far more readily than French ones). So distrust any single number — including ours — and look at the patterns, which are consistent across studies:
| Pattern | Reported range |
|---|---|
| Bounce increase from a dismissible bottom bar | roughly 7–12 points lower than a modal — the cheap option |
| Bounce increase from a full-screen consent wall | studies report 25–40% on affected pages — the expensive option |
| Visitors who leave without choosing anything | about a third in one 2026 study |
| E-commerce acceptance rates | commonly 45–75% with a fair, symmetric banner |
The loss merchants feel is usually measurement, not money
The scariest-sounding stat — "stores lose 20–40% of conversions after adding a banner" — is almost always about reported conversions in analytics, not orders. When a shopper rejects analytics cookies and then buys anyway, Shopify still records the order; your GA4 dashboard just can't see the journey. That gap is real and annoying, but it is not lost revenue, and Google Consent Mode v2 exists precisely to model it back for ad optimization. Before concluding a banner "killed conversion," compare Shopify's own order counts against the analytics report — the orders usually didn't go anywhere.
The four levers that decide what your banner costs
1. Shape. A bottom bar that lets people keep shopping is consistently the cheapest design; a full-screen wall that blocks the product page is consistently the most expensive. No law anywhere requires the wall.
2. Speed. A banner that loads late and shoves the layout around reads as jank and costs trust and Core Web Vitals. The script should be small, deferred, and never render-blocking.
3. Geography. This is the big one for US-heavy stores: no US state requires an EU-style ask-first banner. If you show every visitor worldwide the strict EU treatment, you are paying the opt-in conversion tax on traffic that never required it. Per-region rules — strict where the law demands, light where it doesn't — recover most of the cost. That is the core of how Vafer works.
4. Honesty. Dark patterns (hidden reject, nagging re-prompts) raise acceptance a few points and are exactly what regulators fine — CNIL's penalties against Google and Facebook were for burying "reject." A fair banner with symmetric buttons is both the legal floor and, per the benchmarks above, still yields healthy acceptance in e-commerce.
What we deliberately don't promise
A consent tool that markets "95% acceptance rates!" is selling you dark patterns with a compliance logo on them. We don't publish an acceptance-rate promise, because the honest levers are the four above — and because whether blocking actually works matters more than any acceptance percentage when the auditor shows up. If you want to compare how the consent apps weigh these trade-offs, our comparison of Pandectes, Consentmo and Vafer is written to be checkable.