Collecting and interpreting data about how people actually use a product: which steps they complete, where they drop out, which features are ignored, and what differs between the users who stay and those who leave.
Without numbers, roadmap discussions are decided by whoever argues best. Analytics replaces that with evidence and usually shows that the feature everyone worried about is used by nobody.
We define the handful of events that matter before writing tracking code, set up conversion funnels for the key flows, and keep collection compliant with consent rules so the data is usable in regulated markets.