Google Analytics alternatives that don't need a consent banner

Most people looking for a Google Analytics alternative are not looking for different charts. They are looking to stop showing a banner, stop sending their visitors' data to an advertising company, or stop explaining to a client why the numbers dropped by a third on the day the banner went up.

Those are three different problems, and they narrow the field differently. This is a comparison written around them rather than around feature lists.

What actually removes the banner

Only one property matters: does the tool store anything on the visitor's device, or read anything from it? That distinction is the whole of the consent question, and it splits the market cleanly.

Cookie-based, consent required. Google Analytics 4, Adobe Analytics, Matomo in its default configuration, Microsoft Clarity, and every session-recording tool. GA4 can be configured with consent mode, which models the gaps rather than measuring them — you get estimates for the visitors who declined.

Cookieless by design, no banner needed. Plausible, Fathom, Simple Analytics, Umami, GoatCounter, Matomo configured without cookies, and GhostPane. These compute a daily-rotating identifier server-side instead of storing one on the device.

Within the second group, the remaining questions are about ownership, cost and honesty of the numbers.

Where the data lives, and why it decides the rest

Hosted tools are the pragmatic choice for most sites, and there is nothing wrong with paying someone to run a database. But note what a hosted tool implies: a processor relationship, a data processing agreement, a sub-processor list to keep an eye on, and a subscription that continues as long as you want your own history to remain readable.

Self-hosting removes all four, and adds an operational job in exchange. The size of that job is the real question, and it has changed: running analytics used to mean a server, a database and a backup policy. On an edge platform it can mean a deployment that fits in a free tier and needs no maintenance at all.

What you give up when you drop the identifier

Be honest with yourself about this before switching, because it is the thing people are most surprised by afterwards.

You lose cross-day visitor identity. No "returning visitors over 30 days", no cohort retention curves, no user-level journeys spanning weeks. A person visiting on three days is three daily visitors. If your business genuinely depends on individual-level retention analysis, cookieless analytics is the wrong tool and you should be honest about needing consent for the right one.

You keep almost everything else. Sessions, bounce rate, time on page, entry and exit pages, referrers, campaigns, countries, devices, conversions, revenue per conversion, funnels within a visit, and real-time traffic all survive intact, because they are all computed inside a single day.

Your totals will change. Expect them to move in both directions — the reasons are worth reading before you panic about the first week — and expect the bot filtering to remove traffic your old tool was counting.

The questions to ask a vendor

Feature tables are easy to game. These five questions are not.

1. What exactly is written to the device? The answer should be "nothing", stated without qualification. Anything hedged means a banner.
2. Is the visitor identifier reversible? If the salt is stored anywhere, the identifier is pseudonymous rather than anonymous, and the distinction matters to a regulator.
3. Is the data sampled? Above certain thresholds some tools estimate rather than count. An estimate is fine as long as it says so.
4. Can I export everything, today, without asking? Not "is there an API" — can you get every row out, as a file, in one action.
5. What happens to my history if I stop paying? For a hosted tool, this is a real question with an unpleasant answer more often than you would expect.

Where each option makes sense

Stay on GA4 if you need cross-day user identity, run Google Ads with conversion imports, or have an analytics team already fluent in it. Add the banner, accept the gap, and use consent mode to model it.

Hosted cookieless (Plausible, Fathom, Simple Analytics) if you want the banner gone this afternoon, want someone else to run the database, and are comfortable with a subscription and a processor relationship. This is the right answer for most small sites.

Self-hosted (Umami, Matomo, GoatCounter, GhostPane) if the data must stay in infrastructure you control — a regulated sector, a client who insists, a privacy stance you want to be able to prove rather than assert — or if you run enough sites that per-site subscriptions have become a line item.

Nothing at all is also a legitimate answer, and worth saying: if you look at analytics twice a year and change nothing as a result, a banner and a subscription are both costs without a return. Server logs may be all you need.

Migrating without losing the history

The one thing that stops people switching is twelve months of history they do not want to abandon. It survives: GA4 exports daily totals and per-dimension breakdowns as CSV, and a good importer merges them into the same tables as everything measured afterwards, so a year-over-year chart spans the switch. The full process is a twenty-minute job, and it is worth doing before you turn the old tag off, not after.