Web Analytics Without Google Analytics: What You Gain and What You Give Up

Ask a small team what they use analytics for and the answer, once you strip out the aspiration, is nearly always the same three things. Which pages are people finding. Where did they come from. Did anyone do the thing we wanted them to do.

That is a modest job. It does not need a product built for enterprises buying media at scale, and it certainly does not need a tool you have to relearn every time you open it. The interesting question is not whether the alternatives can answer those three questions — they all can — but what you give up by leaving, and whether you were using any of it.

What GA4 actually costs you

It is free, so the cost shows up elsewhere. Two places, mainly.

The first is retention, and it surprises people. On a standard property, user-level and event-level data can be kept for two months or fourteen, and two months is the default. The longer windows — 26, 38 and 50 months — are available only on Analytics 360. Once data passes the end of its window it is deleted automatically on a monthly basis and what remains is aggregated reporting. If you have ever opened GA4 to compare this quarter against the same quarter last year and found the detail missing, that is why, and it is a setting rather than a bug. Check yours; if nobody ever changed it, you are keeping two months.

The second cost is comprehension. GA4's model of events and parameters is genuinely more flexible than what came before, and that flexibility is the point for a team with an analyst. For a team of four, the flexibility means every simple question requires a custom exploration, and the number you read on Tuesday does not match the one somebody else read on Wednesday. Analytics only helps if people look at it, and people do not look at things that make them feel stupid.

The options, honestly compared

Prices below are published rates at the time of writing and change more often than the products do.

Tool	Cost	Data retention	Best when
Google Analytics 4	Free	Two or fourteen months on a standard property	You buy ads and need the Google integrations
Plausible	From $9/month hosted; community edition self-hostable	Kept for as long as you subscribe	You want one screen everybody can read
Umami	Free self-hosted; hosted tier available	Yours, in your own database	You already run Postgres and a server
Matomo	Free on-premise; hosted from about €29/month for 50,000 hits	Yours, indefinitely	You need the depth without the Google account

The split is cleaner than the table makes it look. Matomo is the full replacement, with the feature list and the corresponding complexity. Plausible and Umami are deliberately narrow: a page of numbers, no configuration, nothing to learn. Which side you belong on depends on whether the last five things you wanted from analytics were on one screen or in an exploration you had to build.

Where consent actually bites

This is the part most comparison articles get slightly wrong, so it is worth being precise about which rule does what.

In the UK, PECR governs storing information on a visitor's device or reading what is already there. Only storage that is strictly necessary to deliver the service the person asked for is exempt, and analytics has never qualified — the ICO's own website classes its analytics cookies as optional rather than essential. That is the rule your banner exists to satisfy. A tool that writes nothing to the device is not caught by it.

Data protection law is a separate matter, and it does not disappear because you dropped the cookie. If a tool processes something that identifies a person, you need a basis for that regardless of where it is stored. What this means in practice is that "cookieless" is a meaningful reduction in the paperwork rather than an exemption from all of it.

Ask a prospective vendor two questions and get the answers in writing: exactly what do you write to the visitor's device, and exactly what do you retain about them. A vendor who cannot answer both in a paragraph is telling you something useful.

There is a practical benefit here that gets overlooked. A consent banner suppresses a meaningful share of your traffic from ever being counted, because a large proportion of visitors decline. Removing the banner does not only simplify the legal position, it makes the numbers more complete — which is a strange argument for the privacy-preserving option, and a true one.

What you actually give up

Three things, and only one of them matters to most small sites.

• Google Ads and Search Console integration. This is the real loss. If you spend money on Google traffic, GA4's attribution is not something the alternatives replicate, and you should keep it. Nothing stops you running both.
• The free export to a data warehouse. Relevant if somebody on your team writes SQL against event data. If nobody does, it is a feature you were describing rather than using.
• Continuity of history. Nothing imports GA4 data cleanly. Accept the gap, export a monthly summary before you leave, and move on.

What you gain is subtler and easier to underrate: a tool people voluntarily open. We have watched teams go from checking analytics when somebody asks a question to checking it in the Monday meeting, purely because the answer is now on the first screen. That change is worth more than any feature on the list above. It is the same pattern behind a weekly publishing workflow that survives contact with a busy week — the tool that gets used is the one with the least friction, not the one with the most capability.

Switching without losing anything

The failure mode is switching on a Friday, deleting the old property, and discovering in March that the only record of last year is gone. A month of overlap prevents it:

1. Install the new tool alongside GA4 and leave both running. Two scripts will not measurably slow your site, and the lighter tools are typically a fraction of the size of the Google tag.
2. Compare a full month. The totals will not match, and that is expected — different bot filtering, different session definitions, and a consent banner suppressing one of them. What matters is that the trends and the top pages agree.
3. Export what you want to keep from GA4 as a spreadsheet: monthly sessions, top pages, top referrers. Not the raw events, which nobody will ever read.
4. Remove the old tag and take the banner down if the new tool genuinely stores nothing.

If you self-host, add one more step: check where the database lives and whether it is included in the backups you already run. An analytics database quietly growing on a server nobody backs up is a small tragedy waiting for a disk failure, and it is exactly the sort of thing a deployment setup you chose for other reasons tends to leave out.

What we would do

If you do not buy traffic, move to a light tool and stop paying attention to numbers you were not using. If you do buy traffic, keep GA4 for the campaign attribution and add a simple tool beside it for everyone else, because the people who need to know which article did well should not have to build an exploration to find out.

Either way, set a calendar reminder to check your GA4 retention setting today. It takes thirty seconds, and two months is a shorter memory than most teams believe they have.