How Matomo works
Matomo is a web analytics platform you can run on your own server. The tracking script sits on your site as usual, but the data it collects lands in a database you control rather than in a vendor’s account. There is also a hosted version run by the company behind it, for people who want the software without the server work. The difference between the two is who holds the data and who patches the machine, not what the reports show.
The reporting will feel familiar to anyone who used the older generation of Google Analytics: visits, pages, referrers, campaigns and goals, organised around sessions rather than around a stream of events. It also offers features that are separate purchases elsewhere, including heatmaps, session recordings, form analytics and A/B testing, though several are premium plugins rather than part of the core install, so check what is included before assuming you have it.
Why Matomo matters
Two reasons, and they are not the same. The first is data ownership. If your organisation handles health, legal or financial information, or if a client contract or a regulator says visitor data must stay in a named country or on your own hardware, self-hosting is sometimes the only arrangement that can be signed off. Nothing leaves your infrastructure, so there is no third-party processor to declare.
The second is unsampled history that you decide the length of. A self-hosted install keeps what you tell it to keep for as long as you are willing to store it, which matters when you need to compare a season with the same season several years back and a hosted product’s retention window will not reach.
Common mistakes with Matomo
Underestimating the running cost is the usual one. The software is free; the server, the backups, the upgrades, the database tuning and the person who does all of it are not. A busy site outgrows a small server quickly, and a Matomo install that has quietly stopped archiving its reports is worse than having no analytics at all, because nobody notices for weeks.
The other is expecting integrations that are not there. Matomo does not plug into Google’s advertising products the way GA4 does, so you will not be feeding Matomo conversions into Google’s bidding, and search performance data arrives through a connector rather than natively. Teams running paid search usually end up keeping GA4 alongside it. That is a reasonable outcome, not a failure of either tool.
What to do about it
Decide the reason first. If the driver is a compliance requirement you can name in a sentence, self-hosting is worth the work and should be planned with whoever looks after your servers. If the driver is general discomfort with Google, ask what would actually change: a hosted, privacy-first analytics service may reach the same outcome with none of the maintenance.
Then settle the boring parts before launch. Where it is hosted, who applies updates, how long raw data is kept, whether IP addresses are anonymised, and exactly how it is described in your privacy policy. Run it alongside your existing setup for a while and expect the totals to differ, because tools count sessions differently and neither one is lying. If what you actually want is a measurement setup that is documented and trusted rather than a particular brand of software, that is the substance of analytics and tracking work.
