Marketers have spent the better part of two decades renting their data infrastructure. You paste a script tag, agree to the terms of service, and in exchange you get a dashboard. What you don't get is the data itself — not really. You get read access to an aggregated view of it, on a platform that can change its interface, its data model, or its pricing at any point. The data lives somewhere else. You've always been a tenant.

This arrangement worked fine when the alternatives were genuinely difficult to operate. That's no longer the case. Tools like umami analytics have made self-hosted, privacy-respecting web measurement practical for organisations that don't have a dedicated infrastructure team — and the case for switching from a rented dashboard to one you actually own is now easier to make than at any previous point. If you want a detailed breakdown of how the setup process works in practice, you can learn more in this guide to running Umami on your own infrastructure. For an overview of what the tool itself offers, the official Umami site is the right starting point.

What is Umami and why are marketers looking at it now?

Umami analytics is an open-source web analytics tool designed to collect the traffic signals marketers actually use — pageviews, referral sources, campaign performance, top content — without cookies, without browser fingerprinting, and without routing visitor data through a third-party platform. Because it runs on infrastructure you control, the data never leaves your environment. You query it directly. You set the retention period. You decide what gets stored and what doesn't.

The timing of its growing adoption among marketing teams is not arbitrary. Consent management requirements in the EU have made cookie-dependent analytics progressively harder to operate cleanly. A properly configured GA4 implementation that includes a consent banner in European markets typically sees between 20 and 40 percent of sessions go untracked — visitors who decline consent or dismiss the prompt before it registers. Umami's cookieless measurement model captures aggregate counts across that audience without triggering the same consent requirements, because no individual visitor identifier is ever created.

That's not a small operational difference for teams that report on audience size, campaign reach, or content performance to clients or leadership.

Umami surfaces the traffic metrics marketing teams need — pages, referrers, campaign tags — without the consent-management overhead that comes with cookie-dependent tools.

The ownership argument is practical, not ideological

There's a version of the self-hosted analytics argument that leans on principle — data sovereignty, surveillance capitalism, the ethics of feeding behavioural profiles to advertising platforms. That argument has its place. But the practical case is compelling enough on its own terms and probably lands better with a marketing budget holder.

Point one: vendor dependency. Rented analytics platforms have a consistent track record of changing what they provide and when. GA4 replaced Universal Analytics on a timeline that gave most organisations insufficient warning to migrate their historical reporting cleanly. Attribution window defaults changed. Metrics were renamed or deprecated. Every time a platform restructures its data model, the marketing team absorbs the cost — in re-instrumenting, in re-training, in rebuilding reports that no longer work. When your analytics runs on infrastructure you control and an open-source codebase you can read, you choose when to update and what to change.

Point two: data access. The data model behind most rented analytics tools is a reporting API, not a database. You can pull what the vendor decided to surface in the interface. With self-hosted umami analytics, visitor event data lands in a Postgres database you own. You can query it with SQL. You can join it against your CRM, your email platform, or your ad spend data without filing a support request or paying for an integration tier.

Point three: consent banner math. For any site with meaningful EU traffic, the percentage of sessions untracked under a well-implemented consent model is not a rounding error. For an editorial site running 200,000 monthly sessions, 25 percent untracked is 50,000 sessions that don't appear in the dashboard at all — with direct implications for advertiser reporting, content planning, and any metric that feeds into business decisions.

Self-hosted analytics means visitor data lives in a database your organisation controls — not behind a reporting API on a platform you don't.

Where the trade-off actually sits

Being direct about what self-hosted umami analytics does not do is more useful than overselling it.

It doesn't replicate the full feature depth of GA4. It won't produce multi-touch attribution paths or integrate with Google Ads bidding signals. For organisations whose marketing stack depends on those connections, umami doesn't replace GA4 — it supplements or substitutes for a specific slice of it. The practical question is which slice you actually use. For the majority of content publishers, B2B sites, and SaaS marketing teams, the answer is: pageviews, referral traffic, campaign performance via UTM parameters, and conversion events. Umami covers that set well.

The operational trade-off is real too. Running a self-hosted instance means someone on your team owns an update cadence, a backup process, and a hosting cost. For small teams, that's typically a VPS in the €5–15 per month range and a thirty-minute update cycle every few weeks. That's a manageable overhead if the alternative is renting measurement infrastructure you can't actually control, query, or extend.

The pattern worth examining in your own stack

Most marketing stacks accumulated their current tools reactively — a new platform launches, someone signs up for the free tier, and six months later it's embedded in three reports nobody wants to rebuild. The result is measurement infrastructure that nobody audited for actual fit, that nobody owns in any meaningful sense, and that will take the team hostage the next time the vendor restructures.

Open-source analytics — and umami analytics specifically — is a useful place to start examining that pattern. Not because every rented tool should be replaced, but because measurement is the one layer where the data genuinely belongs to you, and currently often doesn't.

The stack you rent tells you what the vendor wants you to see. The stack you run tells you what's actually happening.

Cookieless tracking means event data is collected without individual visitor identifiers — no consent banner required in most implementations, and no gap in measurement from users who decline.