The honest answer is: it depends on your banner, and the effect is measurable. A cookie banner can hurt conversions through two separate channels. One is technical, where the banner slows the page or shoves the layout around. The other is behavioral, where the interruption adds friction. A badly built banner damages both. A well-built one costs almost nothing. The difference is engineering and design, not the existence of the banner.
Let's look at what the data actually shows, then how to measure the cost on your own site instead of trusting a generic number.
Channel one: the banner wrecks your page speed
Google ranks and users abandon on Core Web Vitals, and cookie banners are a well-documented source of Web Vitals damage when they're built carelessly.
- Largest Contentful Paint (LCP). If the banner becomes the biggest thing painted on screen, it becomes your LCP element. DebugBear documented a case where a banner pushed LCP from 1.43 seconds to 3.61 seconds, adding over two seconds of delay before the page felt loaded. Render-blocking consent scripts make this worse by holding up everything behind them.
- Cumulative Layout Shift (CLS). This is the classic offender. The page renders, then a fraction of a second later a heavy banner injects at the top or bottom and shoves the content down. The user reaches for a link and it jumps. Late-injected banners are one of the most common causes of layout shift on the web.
- Interaction to Next Paint (INP). When someone clicks accept, a poorly built consent script often runs a long task on the main thread, initializing a pile of third-party tags at once. The page freezes for a beat right at the moment of engagement.
Every one of these correlates with abandonment. A slower, jumpier page loses sales, and the banner is frequently the cause. The fix is entirely within your control, which is the point.
Channel two: friction and design
Beyond speed, the banner is an interruption, and small design choices move behavior more than most teams expect. In a widely cited 2020 study presented at the CHI conference, researchers scraped the consent designs of the top CMPs on popular UK sites and ran a field experiment on how the common patterns affect choices. Two findings stand out. First, only about 11.8% of the banners studied met even the minimal requirements of European law, so most banners on the web are built to nudge, not to inform. Second, small interface decisions, such as moving the notice from the top of the screen to the bottom, substantially changed whether and how people interacted with it. If a design tweak that big changes consent behavior, it changes conversion behavior too, because it's the same interruption sitting between the visitor and your page.
Mobile is where this bites hardest. On a phone, a banner can cover a large share of the visible screen, and mobile shoppers show the least patience for anything between them and the content. A layout that's a minor nuisance on desktop can be a wall on mobile, so always segment your measurement by device rather than reading a blended average that hides the worst case.
You'll see vendor blogs quote conversion drops of 10% to 25% from cookie banners. Treat those as directional, not gospel. They come from individual analyses with different setups, and the range is wide precisely because the effect depends on the banner. The takeaway isn't a magic percentage. It's that the cost is real, variable, and specific to your implementation, which means you should measure yours rather than adopt someone else's number.
Build a banner that doesn't cost you sales
Most of the damage above comes from a handful of avoidable choices:
- Reserve the layout space. Give the banner a fixed slot so it can't shift content when it loads. This kills most CLS on its own.
- Load it async and non-render-blocking. The banner should never hold up the main content paint.
- Keep the script light. A bloated CMP that pulls in megabytes of JavaScript is a tax on every page view.
- Don't cover the fold on mobile, where a banner can swallow a big share of a small screen and where purchase intent is most time-sensitive.
- Keep copy short and the reject option at parity. A clean, honest choice reduces friction and, as it happens, keeps you compliant.
Measure your own banner, don't guess
The right move is to test, not to accept a generic estimate. Watch your field Core Web Vitals in Google's real-user data before and after banner changes. Then run an A/B test that pits banner variants against each other and reads both consent rate and downstream metrics like bounce and conversion, segmented by device. That's how you learn whether your banner is costing you sales and which version costs the least.
CookieBeam is built for exactly this loop. The script is engineered to be lightweight and to avoid layout shift, so channel one is handled by default. The A/B testing tool then lets you run banner variants and compare both consent rates and the conversion and bounce impact, so you optimize for the version that keeps both compliance and revenue. You stop arguing about whether banners hurt conversions in the abstract and start measuring what yours does.
For the deep technical version, read cookie banners and Core Web Vitals. To run the tests, see A/B testing consent banners. And to lift consent without hurting conversions, see raising consent rates without dark patterns and banner copy that increases consent.