Pinpoint your activation metric from candidate list to causal proof

Description

Your team has been guessing at the activation milestone and the retention curve has not budged. This walks the team through brainstorming candidate moments, running a regression against retention, and designing the experiment that proves the milestone is causal, not just correlated.

You are a senior PM identifying the activation metric for {{product_name}}. Current retention curve: {{retention_curve_summary}}. Available data: {{data_sources}}.

## Step 1. Brainstorm candidate aha moments
List 6-10 moments early in the user journey where users plausibly start getting value:
- Account-level moments (workspace created, first invite, first integration)
- User-level actions (first save, first share, first import, first weekly return)
- Workspace-level milestones (3 active editors, 10 items, second seat seated)

For each candidate, write a single sentence describing why a user would hit it.

## Step 2. Filter to actionable candidates
Cut candidates that the growth team cannot move. Activation has to be:
- Observable in product analytics today
- Influenceable by onboarding, in-product nudges, or messaging
- Hit by most users within their first 7 days, not month 3

## Step 3. Run a correlation pass
For each remaining candidate, query 4-week retention split by users who hit it vs users who did not:
- Pick the retention horizon that matches the product (D7 for consumer, W4 for prosumer, M2 for B2B)
- Order candidates by retention lift
- Flag candidates where the lift is at least 2x

## Step 4. Test for the threshold effect
For the top 2-3 candidates, look for the threshold:
- Hitting the moment once vs twice vs five times: does retention keep climbing or plateau?
- Hitting the moment within day 1 vs day 3 vs day 7: at what point does the lift collapse?

The strongest activation metric has a clear plateau at a small repeat count and a clear time bound.

## Step 5. Design the causal experiment
Correlation is not enough. Pick the top candidate and design the test:
- Hypothesis: "If we move the share of users hitting {moment} from X percent to Y percent, retention rises from A percent to B percent."
- Treatment: an onboarding change or in-product nudge that targets the moment
- Sample size and duration that detect the smallest meaningful lift
- Pre-registered guardrails (latency, cost, support tickets)

## Step 6. Stand up the dashboard
Single page that shows:
- Activation rate (percent of new users hitting the moment within the time bound)
- Retention curve split by activated vs not activated
- Trend by week with confidence interval

## Output
1. Candidate list with the why-it-might-be-activation sentence per candidate
2. Filtered shortlist (3-5 candidates)
3. Correlation table (candidate, retention lift, threshold notes)
4. Pre-registered experiment plan for the top candidate
5. Dashboard mockup
6. The runner-up candidate to test next, in case the top one fails the causal test

Customize This Prompt