The job in one line
A growth data analyst answers the question the rest of the team keeps guessing at: is this working, and why. They own the metrics, build the reports, and read the data the way a navigator reads a chart — not for its own sake, but to point the team somewhere. On a growth team, they sit downstream of every other role: the Growth Engineer instruments the events, and the analyst turns those events into evidence.
Retention is the metric under everything
The analyst's first job is usually the retention curve, because retention is the floor every other number stands on. Brian Balfour and the Reforge team argue that weak retention quietly caps acquisition and monetization no matter how good they look. The analyst builds the cohort view — what share of each signup cohort is still active at day 1, 7, and 30 — and watches whether the curve flattens. A flattening curve means the product has a core of users it keeps; a curve that decays to zero means there is no business to scale yet.
From data to decisions
Raw events are not insight. The analyst defines each metric precisely — what counts as active, where the funnel starts and ends, which churn is voluntary — so the team argues about the decision rather than the definition. Then they build the weekly report the Growth PM and Growth Lead read to set the roadmap: the cohort that is growing fastest, the funnel step that leaks, the churn the next month is likely to bring. The output is not a dashboard. It is a recommendation with a number behind it.
Signal versus noise
Most metrics that move do not matter. A good analyst spends as much time killing vanity metrics as building real ones — total signups feel great and predict nothing; activation and retention are harder to move and predict almost everything. The discipline, taught by teams like Amplitude and Reforge, is to find the leading indicator that actually forecasts the outcome you care about, then ignore the rest of the noise that fills a typical dashboard.
How the role usually shows up
On a small team there is rarely a dedicated analyst — the work lands on whoever is most comfortable in SQL, often a founder at 11pm. On a larger team it is a growth analyst or data scientist embedded with the growth function, distinct from a business-intelligence analyst who reports on the company at large. Titles vary; the constant is the same: define the metrics, build the reports, and separate signal from noise so the team bets on evidence.
An analyst on your fromHello team
fromHello runs this role as one of eight specialists in an AI growth team. The Data Analyst agent builds the weekly cohort report, forecasts next month's churn, and flags the segment that is growing fastest — then proposes what to do about it. Nothing ships on its own: the agent proposes, you approve. If you are weighing it against a suite you would still have to staff and run, fromHello vs HubSpot is the honest place to start sizing the difference.