PostHog

PostHog is an open-core platform for product analytics that bundles events, funnels, session recordings and feature flags into one tool. The core is MIT-licensed and self-hostable, while the advanced features live in the hosted cloud.

Where classic web analytics counts how many visitors open a page, product analytics asks what users actually do inside the product: which path they take through an application, at which step they drop off, which feature they never find. PostHog covers this field and at the same time raises the sovereignty question, because the core is available as open source. This page describes what the platform does, where the line runs between the self-hosted core and the cloud, and what privacy angle this carries in Switzerland.

What PostHog covers

PostHog sees itself not as a pure analytics tool but as a collection of several products around a shared event database:

• Product analytics. Events, funnels, retention and user paths. The core: tracing which actions users trigger and where they get stuck.
• Web analytics. Page views, sources and session-level metrics, closer to the classic discipline that Matomo also serves.
• Session recordings. Replay of individual sessions to surface friction points that vanish in aggregate numbers.
• Feature flags and experiments. Switch features on for specific user groups and run A/B tests without a new release.
• Surveys and error tracking. Direct user feedback and runtime errors, tied to the same events.

This breadth is the difference from pure web statistics. It also makes PostHog a larger operational object than a lean analytics script.

Open core, not fully open source

Precision matters here. PostHog is open core, not open source throughout. The bulk of the code is under the MIT license and may be freely used, modified and operated. Whatever sits in the project's ee/ (enterprise edition) directory falls under its own, separate license and is excluded from the MIT release. Anyone self-hosting the core thus gets an open-source foundation, but not automatically every feature of the hosted version. That is a different model from a tool that is open source throughout, and it should be understood before a decision. The underlying trade-off is covered by the page on OSS business models.

Self-hosting: what works and what does not

PostHog provides a Docker Compose variant for self-hosting, but is remarkably candid about its limits. The vendor explicitly describes self-hosting as not officially supported and not recommended for most users; there is no product or infrastructure support for self-operated instances, and the operational risk lies entirely with the operator. Functionally, the self-hosted variant exposes only the features of the free plan; the paid cloud features stay reserved for the hosted platform. So there is no complete free parity with the cloud.

This honesty is the most important property for a sovereignty decision. Self-hosting is possible and the core is open source, but it is a deliberate operational load, not a side switch. Anyone seeking data ownership without that load therefore compares PostHog more with the leaner, fully open-source Matomo.

Cloud variants and data location

In the hosted cloud, PostHog offers a choice between two regions: US (Virginia) and EU (Frankfurt). The region is chosen during setup; an EU region addresses the data-location question for European actors. The separation of layers stays important: the provider is PostHog Inc., based in the US, regardless of which region the data sits in. For Swiss organisations with particularly sensitive data, this distinction between data location and provider jurisdiction is the heart of the trade-off, as also described by digital sovereignty.

The plans follow a usage-based model with a free entry tier. By the vendor's account, the free cloud tier covers around one million events per month for analytics and several thousand session recordings, with shorter data retention than the paid tiers. The exact limits are volume-dependent and should be checked against the current plan before committing.

Product data under the organisation's control

Adopting product analytics is also a decision about how much control the organisation keeps over its event data. A self-hosted core on the organisation's own or Swiss infrastructure keeps the event data within reach; the hosted cloud with an EU region moves the data location to Europe but leaves the US provider in place. Both paths are viable; they differ in the degree of control and the operational effort. Product analytics also almost always touches personal data as soon as sessions can be attributed to individual users. That brings the revised Swiss Data Protection Act (revFADP) into play, and the choice of tool becomes a privacy question.

PostHog thus sits in the same field as other open-source analytics and marketing tools that interact within a sovereign marketing automation setup. The boundary stays clear: PostHog delivers product analytics with the breadth described above, while for pure web statistics without product depth the leaner, fully open-source Matomo is the better-fitting tool.

References

• PostHog Pricing. Usage-based plan model with a free entry tier, including the free allowances for events and session recordings. (2026). posthog.com/pricing
• PostHog Self-hosting documentation. Docker Compose variant for self-hosting, with the explicit note that it is not officially supported and not recommended for most users. (2026). posthog.com/docs/self-host
• PostHog LICENSE. MIT license for the core, separate license for the ee/ (enterprise edition) directory. (2026). github.com/PostHog/posthog/blob/master/LICENSE

---

Related topics

• Matomo, the leaner, fully open-source alternative for web analytics.
• Digital sovereignty, the trade-off between data location and provider jurisdiction.
• Marketing automation, the setting in which analytics tools interact.
• nFADP / revFADP, the Swiss data protection law that applies to user data.
• OSS business models, the logic behind the open-core model.