Analytics & CRO
Running Matomo on-premise was a deliberate decision, not a default. Most analytics teams inherit whatever tool is already in place — GA4, a SaaS platform, whatever the previous team set up. I made the switch intentionally, and three years later it is still the right call for the kind of work I do.
This is a practical overview of what Matomo on-premise actually involves: the architecture, what it costs in real terms, what you get that a SaaS analytics tool doesn’t, and where the tradeoffs are.
What on-premise actually means
Matomo on-premise is a self-hosted analytics platform. You download the software, run it on your own server, and all data stays within your infrastructure. No data leaves for Matomo’s servers. No third-party processing. No external dependencies on the data itself.
That is the key distinction from Matomo Cloud: with on-premise, Matomo provides the software and you run the infrastructure. With Cloud, Matomo handles both. The tracking code is the same. The reports are largely the same. The fundamental difference is where the data lives and who can access it.
Why data ownership matters in practice
For an online travel agency operating across European markets, GDPR isn’t a compliance checkbox — it’s a constraint that shapes every analytics decision. The question I kept running into with GA4 wasn’t whether we were compliant: it was whether we could prove it. Data processed on Google’s servers, subject to US law under the Cloud Act, in a regulatory environment where Schrems II had invalidated the previous adequacy framework — the legal team was uncomfortable, and so was I.
Matomo on-premise removed that uncertainty. The data doesn’t leave our infrastructure. We can configure tracking to run without cookies by default, which means most users don’t trigger a consent requirement at all. We produce clean, unsampled data I can present in a product review without caveating what might be missing due to consent drop-off.
That’s not an argument that applies universally. If you’re running a US-only SaaS with no EU exposure, GA4 probably makes more sense. For a travel platform with a European customer base and a legal team paying attention, the case is clear. For a full platform comparison, see Matomo vs GA4: what actually changes.
What the setup actually involves
Getting Matomo on-premise running is a one-time engineering task: server provisioning, database setup, software installation, and placing the tracking snippet on the site. Once it’s up, day-to-day operation requires minimal maintenance — mostly version updates and occasional database optimisation as data volume grows.
The ongoing operational overhead is lower than most teams expect. The tracking snippet goes on the site via Matomo Tag Manager, which means tracking changes don’t require engineering involvement. Reports, goals, and segments are configured in the Matomo interface by whoever manages the analytics function. Our engineering team hasn’t touched the Matomo installation in over a year, outside of version upgrades.
What you get out of the box
The core installation covers the fundamentals well: visit tracking, goal conversion tracking, e-commerce reporting, campaign tracking, and segmentation. For most analytics use cases, that’s sufficient.
The significant advantage over GA4 is data fidelity. Matomo records every session, not a statistical sample. For a site doing hundreds of thousands of sessions per month, sampled reports create doubt about whether you’re seeing the real numbers or an extrapolation — and that doubt shows up in product reviews when someone challenges the data.
Where the core installation falls short is behavioural depth: there’s no heatmapping, session recording, A/B testing, or form analytics in the free version. Those require premium plugins, which I cover in Matomo’s premium plugins: what you actually get. For a team coming from Hotjar plus GA4, the premium bundle is roughly equivalent in cost — with data ownership on top.
The tradeoffs worth knowing
On-premise isn’t the right choice for every team.
You need someone with server access for installation and maintenance. Running an outdated Matomo installation is a security risk, so version upgrades need to be on someone’s radar. If your analytics needs are simple and your team is small, the overhead probably isn’t worth it compared to Matomo Cloud.
Matomo also doesn’t match GA4 on integrations. If your marketing stack runs through Google — Ads, Looker Studio, BigQuery — GA4 connects natively. Matomo integrations exist but require more configuration. That’s worth factoring in before making the switch.
On-premise or Cloud? If GDPR compliance and data ownership are the driving reasons, on-premise gives you more control. If you want Matomo’s privacy model without the server overhead, Matomo Cloud gives you most of the same reporting with Matomo handling the infrastructure. The data stays in Europe either way — the difference is operational.
The practical test: can you confidently answer “where is our analytics data stored, and who can access it?” for your current tool? If not, that’s where the conversation should start.
