The job in one line
A growth PM decides what to test next. They take the direction set by the Growth Lead — the metric, the strategy — and turn it into a concrete experiment roadmap: a ranked list of tests, each with a hypothesis and a clear win condition. They are one of the two steering roles on a growth team, sitting between the growth lead who sets the metric and the marketers, engineers, and analysts who run the tests.
The test brief is the core artifact
The PM's first move is to make a vague goal testable — "improve activation" becomes specific bets like shorten the signup form or add a setup checklist. Each bet becomes a brief, the deliverable a growth PM is judged on. A good one states the hypothesis (what we believe and why), the primary metric, the variants, the audience and sample, and what would count as a win or a loss before the test runs. Written well, the brief lets a marketer or engineer build the test without a meeting — and lets the analyst read the result without re-litigating the setup.
Prioritization: impact, confidence, effort
More ideas exist than the team can run, so the PM ranks them. The common shorthand is ICE — impact, confidence, ease — popularized by Sean Ellis and widely documented by teams like those writing at Reforge. RICE adds reach to the front. The framework matters less than the discipline: score every idea on the same axes, and the highest-expected-value bet goes first.
Cadence, not channels
A growth PM rarely runs a channel themselves. Their output is the rate of validated learning: how many well-specified tests ship per cycle, and how cleanly each one resolves. Brian Balfour frames growth as a repeatable process — goal, ideas, prioritize, test, analyze, run on a tempo. The PM owns that tempo and keeps the backlog moving, while the rest of the team executes.
How fromHello runs this role
fromHello runs the growth PM as one of eight roles on an AI growth team. The agent takes the metric from the Growth Lead, generates hypotheses against your funnel, scores the backlog, and drafts the briefs — then proposes them for you to approve or edit. It is pre-launch and human-in-the-loop by default: nothing ships without your sign-off. For sizing against a traditional suite, see fromHello vs HubSpot.