Fathom Analytics vs Umami
Fathom Analytics and Umami both answer the same question: how do you know what users actually do? Privacy-first simple analytics with strong EU data isolation and a bypass for ad blockers via your own domain. Open source privacy-focused analytics you can self-host on your own database, or use as a hosted service. The real split is ownership: Umami runs inside your project and leaves the operational work with you, while Fathom Analytics runs the hard parts as a service and takes a dependency in exchange.
Verdict
The real split is ownership: Umami runs inside your project and leaves the operational work with you, while Fathom Analytics runs the hard parts as a service and takes a dependency in exchange.
Pick Fathom Analytics if
- ✓Unlimited websites on one subscription suits people running many small projects.
- ✓Custom domain proxying recovers traffic that ad blockers would otherwise hide.
- ✓EU isolation option keeps European visitor data out of other jurisdictions.
Pick Umami if
- ✓Self-hosting means the raw analytics data sits in a database you already run.
- ✓Stores into Postgres or MySQL, so you can query it with plain SQL.
- ✓No licence cost at all when you supply the hosting yourself.
| Comparison | Fathom Analytics | Umami |
|---|---|---|
| Pricing shape | Paid monthly plans by page view volume, with unlimited sites included on every tier. | Free if you self-host on your own infrastructure, with a paid cloud plan by event volume. |
| Frameworks | Next.js, SvelteKit, Nuxt, Laravel, Django, Rails | Next.js, SvelteKit, Nuxt, Laravel, Django, Rails |
| In one line | Privacy-first simple analytics with strong EU data isolation and a bypass for ad blockers via your own domain. | Open source privacy-focused analytics you can self-host on your own database, or use as a hosted service. |
Pricing described qualitatively because published plans change often. Checked 2026-08-23. Confirm current terms on Fathom Analytics and Umami.
Fathom Analytics
Strengths
- ✓Unlimited websites on one subscription suits people running many small projects.
- ✓Custom domain proxying recovers traffic that ad blockers would otherwise hide.
- ✓EU isolation option keeps European visitor data out of other jurisdictions.
- ✓Uptime monitoring is included rather than sold as a separate product.
Tradeoffs
- ✗Closed source, so you cannot audit or self-host the collection code.
- ✗Reporting stays deliberately shallow, with no funnels or cohort analysis.
- ✗Data export options are limited if you want your own warehouse copy.
- ✗Page-view pricing gets expensive for high-traffic content sites.
Umami
Strengths
- ✓Self-hosting means the raw analytics data sits in a database you already run.
- ✓Stores into Postgres or MySQL, so you can query it with plain SQL.
- ✓No licence cost at all when you supply the hosting yourself.
- ✓Multi-site and team support without paying per additional property.
Tradeoffs
- ✗Self-hosting means you own uptime, upgrades and database growth forever.
- ✗Analytics traffic writes to a database you must size and prune yourself.
- ✗Feature set trails commercial tools, especially around alerting and reporting.
- ✗Community support only, so an obscure bug can stall you.
Frequently asked questions
Is Fathom Analytics or Umami better?
Neither is better in the abstract. The real split is ownership: Umami runs inside your project and leaves the operational work with you, while Fathom Analytics runs the hard parts as a service and takes a dependency in exchange. Pick the one whose downside you can absorb, because both upsides are real.
What is the main drawback of Fathom Analytics?
Closed source, so you cannot audit or self-host the collection code. Reporting stays deliberately shallow, with no funnels or cohort analysis.
What is the main drawback of Umami?
Self-hosting means you own uptime, upgrades and database growth forever. Analytics traffic writes to a database you must size and prune yourself.
Can you switch from one to the other later?
Usually, at a cost that grows with how much of your product leans on the analytics layer. Plan for it in the data model, not in the framework, and the switch stays survivable.