Here is an uncomfortable number: on many European sites, the majority of visitors never appear in Google Analytics. Not because they were not there, but because they declined the cookie banner, or ignored it, and GA4 only counts people who opted in. You are making decisions about traffic, campaigns and conversion on a sample that is both smaller than you think and skewed toward the users most comfortable being tracked.
Meanwhile the legal ground keeps shifting under GA4. Several European data protection authorities have ruled specific Google Analytics setups unlawful over transfers of personal data to the US, and GDPR enforcement has only accelerated, with cumulative fines now well past ten billion euros. For a business operating in the EU, "everyone just uses Google Analytics" has stopped being the safe default it once was.
Cookieless, privacy-first analytics is the response a lot of teams are landing on. It measures what matters without setting tracking cookies, which in turn means no consent banner for analytics, no lost data, and a much shorter conversation with your DPO. Here is how it actually works and how to move without losing your history.
Why the cookie banner is costing you data
The consent banner is not just an annoyance, it is a measurement bug. When analytics depends on cookies that require consent, every visitor who says no, or closes the banner without answering, becomes invisible. Opt-in rates vary wildly by audience and design, but a large slice of your traffic routinely goes uncounted.
The damage is not evenly spread, which is the real problem. Privacy-conscious users, often your more technical or higher-intent visitors, decline at higher rates. So you are not just losing volume, you are losing a biased slice of it. Your conversion rate looks wrong, your channel attribution looks wrong, and you cannot tell by how much. A tool that counts everyone, even approximately, beats a tool that counts a self-selected subset precisely.
Precise but biased vs approximate but complete
GA4 gives you very detailed data about the people who consented. A cookieless tool gives you slightly less detail about everyone. For most product and marketing decisions, the second is more useful, because a trend you can trust beats a decimal place you cannot.
How cookieless analytics actually works
The obvious question is: if you are not setting a cookie, how do you know a returning visitor from a new one? The honest answer is that you approximate it, and you do so in a way that stores nothing personal.
A genuinely cookieless tool sets zero cookies of any kind. Instead of a persistent identifier, it derives a temporary, rotating hash from signals like the IP address, the user agent and the day, salted so it cannot be reversed. That hash lets it recognise the same visitor within a single day, then it is gone. There is no cross-site profile, no persistent ID, and nothing that ties a session back to a named person. Because no personal data is stored and no device is tracked across visits, most privacy regulators treat it as not requiring consent at all.
You lose some things. Precise returning-visitor tracking over weeks, individual user journeys stitched across sessions, and some remarketing hooks all get weaker or disappear. What you keep is the stuff most sites actually use daily: page views, top pages, referrers and channels, country, device type, and conversion events like a form submit or a signup.
What you can and cannot measure
It helps to be blunt about the trade before you switch, so nobody discovers a missing report during a board meeting.
Works well without cookies:
Traffic volume, trends and top content
Where visitors come from: search, referral, direct, campaign UTMs
Conversions and goal completion (contact form, signup, checkout)
Country, browser and device breakdowns
Bounce and engagement at the page level
Gets harder or goes away:
Deterministic cross-session user journeys
Long-window cohort and retention analysis tied to individuals
Some ad-platform remarketing audiences
Granular demographics that Google infers from account data
For a marketing site, a blog or most SaaS funnels, that first list is 90% of what you look at. If you genuinely run sophisticated multi-touch attribution, you may keep a consented tool for a subset of users and put a cookieless tool underneath it for the complete picture.
Choosing a tool
The category is crowded now, which is good news. The names that come up repeatedly are Plausible, Fathom, Matomo, Umami, PostHog, Simple Analytics and Pirsch. They split roughly into three groups.
TypeExamplesBest whenLightweight hostedPlausible, Fathom, Simple AnalyticsYou want GA-style dashboards, fast setup, no maintenanceSelf-hosted / open sourceUmami, Matomo, Plausible CEYou want to own the data on your own infrastructureProduct analyticsPostHogYou need funnels, session replay and feature flags too
A few things to weigh beyond the feature list: where the vendor hosts data (an EU-hosted provider sidesteps the transfer question entirely), whether the script is light enough not to hurt your Core Web Vitals, and whether it is genuinely cookieless or just "cookie-optional". Read the privacy page, not the landing page.
Run both for a month
Do not cut over blind. Install the new tool alongside GA4 for a few weeks. You will usually see the cookieless tool report meaningfully more sessions, because it is counting the people your banner was hiding. That gap is the data you were missing, and it is the argument that ends the internal debate.
Migrating without losing your history
The switch itself is easy. The planning is where teams trip up.
Export first. Whatever you keep from GA4, pull it out before you reduce your reliance on it, because Google does not keep old data around forever and you cannot backfill it later. A yearly export of your key reports is usually enough.
Then decide what a conversion is, in plain language, before you configure anything. Most cookieless tools track events explicitly rather than inferring them, so "a lead" has to be defined as a specific action, a submit on the contact form, a signup completion, before it shows up in a dashboard. Writing those definitions down is a good exercise regardless of tooling.
Keep a short overlap period where both tools run, reconcile the numbers so you understand the new baseline, and only then remove the analytics cookie from your consent banner. If analytics was the main thing your banner was gating, you may be able to simplify or remove the banner entirely, which is a small win your visitors will feel on every single page.
None of this requires a big project. It requires deciding that measuring everyone approximately, without a banner and without a legal grey area, is worth slightly coarser data. For most businesses operating in Europe in 2026, it plainly is.
