What product analytics actually measures
Product analytics answers behavioral questions: which steps convert, where people drop off, whether they come back
a week later, and what they do next. It works on events — discrete actions like a page view, a
click, or a custom signup — each tied to a person. That’s the key difference from
web analytics, which counts traffic and sources at the page
level rather than tracking behavior per person over time.
Why “open source” matters here
Behavioral data is sensitive — it’s a detailed record of what your users do. Open source product analytics lets you self-host the whole tool, so those events live on your own infrastructure instead of a vendor’s cloud. You can read the code, audit how data is processed, and avoid both lock-in and per-event SaaS pricing. For many teams that’s the entire reason to choose open source over a proprietary product like Mixpanel or Amplitude.
How it works: events and identity
Two ideas do most of the work. First, event capture: a small SDK sends events as they happen.
Good tools autocapture common interactions — page views, clicks, scrolls, form submits — so you don’t have
to instrument everything by hand. Second, identity resolution: before sign-in, events belong to an
anonymous ID; when you call identify(userId), that history merges into a single profile. Traits like
plan or email then live on the profile and filter every report.
Under the hood, events are usually written to a columnar database built for analytical queries. That’s what lets a funnel or retention report scan millions of events quickly. You can read more about one such event pipeline and how the pieces fit together.
The core insight types
A capable product-analytics tool ships these out of the box:
- Trends — a time series of any event, counted or aggregated, broken down by properties.
- Funnels — conversion across ordered steps, with drop-off and time between steps.
- Retention — cohort heatmaps showing who comes back, and when.
- Segmentation — one metric sliced by any dimension, for KPI tiles and scorecards.
- User flows — a Sankey of the paths people actually take between events.
- Top K — ranked values of a property: top pages, referrers, or products.
You can see these in practice on the analytics overview.
How to choose an open source product analytics tool
Weigh three things. Scope — do you only need analytics, or also session replay and feature flags (in which case a broader platform fits)? Operational cost — some tools are a single binary, others a multi-service stack; be honest about what your team can run. Data ownership — confirm you can self-host and export freely. If you specifically need privacy and data residency, our privacy-first analytics guide goes deeper, and the self-hosted analytics page covers running it yourself.