Analytics glossary. Product analytics terms explained.

Measuring how people actually use a product, event by event, to answer questions like what makes users stick, where they drop off, and what to fix, using funnels, retention, cohorts, and paths rather than raw pageview counts.

An analytics backend that speaks the Model Context Protocol, so a coding agent like Cursor or Claude can query your real numbers in plain English and get answers computed from your data, not guessed, in the same window where you write code.

Tracking the share of users who complete each step of an ordered flow (say signup, activate, pay) to find the single step where the most people drop off, so you know exactly which part of the journey to fix first.

The share of users who come back and keep using a product over time, measured as day-1, day-7, or day-N retention, so you can see whether new users stick or quietly leave after the first visit.

Grouping users by something they share (the week they signed up, a first action) and tracking each group over time, so you can compare how different cohorts retain or convert instead of averaging everyone together.

The number of unique users who did something meaningful in the last 30 days, the standard measure of active usage, often paired with DAU and the DAU/MAU ratio as a stickiness signal.

Recording the specific actions people take (a signup, a click, a purchase) as named events with properties, so you can build funnels, retention, and paths instead of only counting pageviews.

Measuring visitors and usage without storing identifiers on the device, so there is no cookie banner and no GDPR consent step, while still counting visitors, referrers, funnels, and retention.

Deciding which channels and campaigns get credit for a signup or purchase, using UTM parameters and first-touch or last-touch rules, so you know which sources actually drive results instead of guessing.

Looking at the actual sequences of actions users take (what they do after signup, where they wander before converting) as observed, undirected journeys, rather than the fixed ordered steps of a funnel.

The share of users who complete a desired action out of everyone who had the chance, measured per funnel step or end to end, so you can quantify how well a flow turns visitors into the outcome you want.

Analytics that records interactions (pageviews, clicks, scroll depth, form submits, errors) automatically from one snippet, no per-event tracking code. It keeps the element behind each click, so you can name any button as an event later, retroactive to install.

The work of making a product record user actions as named events with properties, so a funnel or retention report has clean data behind it. Done by hand, retroactively from captured clicks, or written and verified for you by a coding agent.

The single agreed spec of every event you record: each name, its properties, and meaning. As plan-as-code it lives in the repo and a checker fails the build when code or traffic drifts from it, so event names can't silently splinter.

A burst of rapid clicks on the same spot in a short window, usually three or more in about two seconds. It signals a user expected something to happen and it didn't, flagging a broken, slow, or misleading control without needing a survey.