Web & product analytics: who comes and what people do

Two different questions, two families of tools

People often confuse "web analytics" and "product analytics", yet they answer distinct questions. Web analytics measures traffic: how many people arrive, where they come from, which pages they land on and which convert. Product analytics measures usage: inside your application or service, which actions users perform, in what order, and which ones make them stay.

A brochure site or a blog mostly needs web analytics. An application, a SaaS, or an interactive tool needs both. Choosing the right tool starts with knowing which of these two questions is burning hottest for you.

Google Analytics 4: the free reference, but demanding

Google Analytics 4 (GA4) is the most widespread tool, free, and integrated into the Google ecosystem (Search Console, Google Ads, Looker Studio). It measures traffic, acquisition sources, conversions, and a bit of product analytics through its event-based model.

Its strengths: it's free, deep, and connects natively to advertising tools if you run paid acquisition. Its weaknesses: an interface reputed to be difficult, a data model that's confusing for beginners, and above all compliance stakes — GA4 sends data to Google's servers, which in Europe requires a correctly configured consent banner (see the tracking & GDPR chapter). Without consent, part of your traffic simply isn't measured.

GA4 is the right choice if you run Google ads, if you need free depth, and if you accept a learning curve.

The privacy-friendly alternatives

A family of tools has emerged to address GA4's heaviness and GDPR constraints: simple, lightweight, often cookieless analytics — therefore sometimes usable without a consent banner.

Plausible: open source, hosted in Europe, cookieless, a single screen showing the essentials (visitors, sources, pages, conversions). Indicative price: from around €9/month. Ideal for a brochure site, a blog, or a SaaS that wants readable numbers without a sprawling machine.
Fathom Analytics: same philosophy, paid, focused on simplicity and privacy.
Matomo: open source, installable on your own server (so your data stays with you), more complete than Plausible but also heavier. Paid cloud version or free self-hosting.
Microsoft Clarity: free, no limit, but mainly a behavior tool (heatmaps, recordings) — see the next chapter.

The trade-off is clear: these tools give less depth than GA4, but immediate readability and simpler compliance.

Product analytics: Mixpanel, Amplitude, PostHog

If you have an application where people do things (sign up, create a project, invite a teammate, go premium), web analytics isn't enough. You need to track events and analyze journeys and cohorts.

PostHog: open source, all-in-one (product analytics + heatmaps + session recordings + feature flags + A/B testing). Generous free plan. It's often the best starting point for a solo SaaS, since it bundles several of this program's bricks into one tool.
Mixpanel: a product-analytics specialist (funnels, retention, cohorts). Decent free plan, then paid by event volume.
Amplitude: very powerful on behavioral analysis and retention, free plan, but can become complex and expensive at scale.

The common principle: you define the events that matter (for example signup, project_created, upgrade_premium), you send them to the tool, and you can then answer questions like "what percentage of sign-ups create a project within 7 days?".

How to choose, concretely

graph TD
    A[What is my need?] --> B{Brochure site or blog?}
    B -->|Yes| C[Plausible or GA4]
    A --> D{Application / SaaS?}
    D -->|Yes| E[PostHog or Mixpanel]
    A --> F{Google ads / SEA?}
    F -->|Yes| G[GA4 essential]
    A --> H{Privacy priority?}
    H -->|Yes| I[Plausible, self-hosted Matomo]

Practical starting rule: a simple site → Plausible (readable, compliant) or GA4 if you run ads. An application → PostHog (all-in-one, free to start). You don't take both families "just in case": you install the one that answers the current question, cleanly, and add the other when the need genuinely appears.

The "I'll install it and look later" trap

Installing an analytics tool is pointless if no one defines what it should measure or ever looks at it. Many entrepreneurs have had GA4 on their site for months without ever opening the report. Before installing, write the three questions the tool must answer (for example: "where do my buying customers come from?", "which page loses the most visitors?"). That's what turns a tracking script into a decision system.