Analytics that actually drive decisions
Why every product site needs its own funnel, the 6 events that cover 80% of insights, and why we don't trust an aggregate conversion rate.
A site without analytics is a shop where nobody counts customers. Here is how we set up analytics for every new landing page and why we never optimise for a single "overall conversion".
The 6 events we always track
Hero CTA, sticky CTA, contact form, scroll to 80%, case-study click, pricing click. That is enough to see the problem without drowning in data. Anything else we add only when a hypothesis demands it.
Why aggregate conversion lies
An overall conversion rate hides the most important fork: did the user come from an ICP channel or random traffic? We always split the funnel by at least three sources and judge them separately. It saves us from optimising things that were never broken.
Where the events live
GA4 and Plausible for traffic and attribution. PostHog for the product funnel and session replays. UTM tags carry campaign and creative — without them, two months later nothing tells you what worked.
What we do with the numbers
Once every two weeks we review the dashboard with the client: one hypothesis, one metric, one decision. More than two hypotheses on the table is no longer analytics, it's a ritual.