Cookieless analytics: Plausible and Fathom compared to GA4

The first reason is consent. Google Analytics uses cookies and processes data that can identify a device, so in the EU it needs opt in consent before it runs. Typical acceptance rates sit somewhere between forty and seventy per cent depending on the audience and how the banner is built.

That means the number in your dashboard is not your traffic, it is the traffic of the people who clicked accept. Trends stay usable, absolute numbers do not.

The second reason is complexity. GA4 is an event based system designed for large advertisers, and a small business owner who wants to know how many enquiries came from Google has to learn a reporting model built for a different problem.

The third is data retention. GA4 keeps user level data for a limited window by default, so year over year comparisons need either an export or a paid setup.

The counter argument is equally real. GA4 is free, it integrates natively with Google Ads and Search Console, and every marketing agency you might hire already knows it.

Cookieless tools identify a visit without storing anything on the device. Instead of a cookie they generate a daily hash from the IP address, the user agent and a rotating salt, then throw the raw inputs away.

The practical consequence is that a visitor is recognised within a day but not across days. If somebody reads your article on Monday and returns on Thursday, that is two visitors, not one returning visitor.

This is what removes the consent obligation in most interpretations, because no data is stored on the device and no personal data is retained afterwards. That said, the legal position depends on national guidance, so treat it as a strong simplification rather than a blanket exemption.

There is a second benefit that surprises people: cookieless scripts are tiny. Plausible ships around one kilobyte against roughly fifty for the Google tag, and that difference is visible in Core Web Vitals on mobile.

Ad blockers matter too. A significant share of technical audiences block the Google tag entirely, so GA4 undercounts them even when consent is given, while self hosted or first party analytics endpoints are usually not blocked.

Be honest about the trade before switching, because some of the losses are permanent.

• •You keep pageviews, sources and referrers, entry and exit pages, countries, devices, browsers, and custom goals such as form submissions or clicks on a phone number.
• •You keep campaign tracking, because UTM parameters work exactly the same way.
• •You lose cross session user journeys, so you cannot see that a person visited four times over two weeks before buying.
• •You lose demographic and interest data, which came from Google's advertising profile rather than from your site.
• •You lose native Google Ads integration, so conversion imports back into Ads need a separate mechanism.
• •You lose deep funnel exploration and cohort analysis, tools that most small sites open once and never use again.

For a service business that measures enquiries, almost nothing on the loss list is used in practice. For an online store running paid acquisition, the Google Ads link alone can justify keeping GA4.

Plausible is the best known choice. It is open source, EU hosted in Germany, and costs roughly nine to twelve euros a month for ten thousand pageviews. You can also self host it for free if you are comfortable running a container.

Fathom is very similar in spirit, priced comparably, with EU data processing available. The difference between the two comes down to interface preference more than capability.

Matomo is the heavyweight. It is the closest thing to a full GA4 replacement with funnels, heatmaps and session recording, it can run on your own server, and its cloud version costs meaningfully more. It can be configured either with or without cookies.

Umami and Counter are lighter open source options for people who already have a server and want zero recurring cost.

Simple Analytics and Pirsch round out the field with the same privacy first model and a small monthly fee.

One combination worth considering: run a cookieless tool as your everyday dashboard and keep Search Console for search data. Search Console is free, needs no consent, and already tells you which queries bring clicks.

Choose by what decision the data will drive. If you are not running paid campaigns and your key metric is enquiries or orders, a cookieless tool covers you completely and removes a banner problem at the same time.

If Google Ads is a real channel with a real budget, keep GA4 for the conversion loop and add a cookieless tool alongside it for honest totals. Running both is normal and costs little.

Switch at the start of a month and run both systems in parallel for four to six weeks. Your new numbers will look higher, sometimes by half, because you are finally counting the visitors who declined the banner.

Export your historical GA4 data before you stop using it. Free GA4 retention is limited, so a one time export of monthly totals for traffic, sources and conversions protects your ability to compare years later.

Do not remove the consent banner just because analytics no longer needs it. If you still load Meta Pixel, Google Ads tags, YouTube embeds or Hotjar, consent is still required for those.

Finally, set up goals on day one. An analytics tool without a defined conversion is a vanity dashboard, and the entire point of switching is to get numbers you can act on.