Skip to content

SKILL.md

Latest commit

 

History

History
92 lines (67 loc) · 2.57 KB

SKILL.md

File metadata and controls

92 lines (67 loc) · 2.57 KB
name growth-experimentation-advisor
description Design a structured growth experimentation program — identifying levers, generating hypotheses, designing cheap tests, and building a culture of data-driven experimentation.
type interactive
domain growth
difficulty advanced
estimated_time 25-35 min
prerequisites
growth-model
ab-test-design
outputs
experimentation-roadmap
triggers
Build a growth experimentation program
What experiments should we run
Growth experiment pipeline
Experimentation framework

Purpose

Design a structured growth experimentation program — identifying high-leverage growth levers, generating hypotheses, designing cheap tests, and building a backlog of experiments to run systematically.

Application

Turn 1: Growth Levers

Agent asks: "Where in your AARRR funnel is the biggest opportunity?"

Agent reviews:

  • Current metrics per stage (acquisition, activation, retention, revenue, referral)
  • Biggest drop-off points
  • Most impactful lever (usually activation or retention)

Turn 2: Hypothesis Generation

Agent asks: "For the [identified lever], let's brainstorm 5+ experiment ideas."

For each idea:

**Hypothesis:** If we [change], then [metric] will [improve by X%]
**because** [mechanism/evidence]
**Effort:** [S/M/L]
**Expected impact:** [H/M/L]

Turn 3: Prioritize Experiments

## Experiment Backlog (Prioritized)

| # | Hypothesis | Lever | Impact | Effort | ICE Score | Status |
|---|-----------|-------|--------|--------|-----------|--------|
| 1 | [Hyp] | [Lever] | [H/M/L] | [S/M/L] | [#] | [Queue/Running/Done] |
| 2 | [Hyp] | [Lever] | [H/M/L] | [S/M/L] | [#] | [Status] |

Turn 4: Experiment Design (Top 1-2)

Agent generates detailed experiment designs using skills/growth/ab-test-design/SKILL.md

Output

# Growth Experimentation Roadmap

## Current Quarter Focus
**Primary lever:** [Activation/Retention/etc.]
**Target metric:** [X → Y in Z weeks]
**Experiment velocity:** [# experiments per sprint]

## Experiment Pipeline
[Prioritized backlog from Turn 3]

## Cadence
- Weekly: Review running experiments
- Biweekly: Launch new experiments
- Monthly: Update growth model with learnings

References

Related Skills

  • skills/growth/ab-test-design/SKILL.md — Design individual experiments
  • skills/growth/growth-model/SKILL.md — Growth model drives lever selection
  • skills/growth/activation-optimization/SKILL.md — Activation experiments

Skill type: Interactive Domain: Growth & Optimization