Installs into .claude/skills of the current project.
Are you the author of Growth Hacker?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/aicodedecode-growth-hacker)
---
name: growth-hacker
description: Run growth experiments — acquisition loops, activation, retention, and referral systems driven by rapid testing.
category: business-marketing
---
## Overview
Growth hacking is marketing with an experimental mindset: identify the highest-leverage growth opportunities, run fast cheap tests, double down on winners, and kill losers quickly. This skill covers the growth process (ICE scoring, experiment sprints), the AARRR funnel, growth loops, and the core levers: acquisition, activation, retention, referral, revenue.
It's a complement to, not a replacement for, brand and product fundamentals — growth tactics amplify a good product; they can't save a bad one.
## When to use
- Growing a product or service with limited budget
- Improving activation or onboarding conversion
- Designing referral or viral loops
- Reducing churn
- Prioritizing among dozens of growth ideas
- Setting up a growth experimentation cadence
- Prioritizing a backlog of 50+ growth ideas
- Designing onboarding experiments to lift activation
- Building a growth team or growth process from scratch
- Finding product-channel fit for a new product
- Breaking through a growth plateau
- Building viral or referral mechanics
## Core concepts
**AARRR funnel.** Acquisition → Activation → Retention → Referral → Revenue. Diagnose which stage leaks most, and focus experiments there. Most teams over-invest in acquisition while activation bleeds.
**Growth loops over funnels.** Loops compound: a user action creates value that brings in the next user (viral loops: sharing; content loops: UGC → SEO → users; sales loops: revenue → more sales capacity). Design loops, don't just pour into funnels.
**North Star Metric.** One metric capturing core product value delivered (e.g., messages sent, projects completed). Every experiment should plausibly move it. Supporting input metrics track the levers.
**ICE scoring.** Rate ideas on Impact, Confidence, Ease (1–10 each); prioritize highest total. Keeps the backlog objective and prevents HIPPO-driven priorities.
**Experiment discipline.** One hypothesis, one primary metric, minimum sample size, fixed runtime. Document learnings win or lose — a failed test that teaches is valuable.
**Activation.** The moment a new user experiences core value ("aha moment"). Shorten time-to-value: reduce signup friction, guide with onboarding, personalize the first session. Activation rate is often the highest-leverage metric.
**Retention curves.** Plot retention by cohort over time. Healthy products flatten; leaky ones decay to zero. Segment by acquisition channel and behavior to find what retains.
**Compounding vs. one-time wins.** Prefer mechanisms that compound (loops, SEO content, referral systems) over one-time spikes (PR hits, launch-day traffic). Score ideas not just on immediate impact but on whether they build durable advantage.
**The experiment review ritual.** Weekly 30-minute review: what shipped, what the data says, what we learned, what is next. The ritual matters more than any single test — it creates the cadence that compounds.
**Pirate metrics (AARRR).** Acquisition → Activation → Retention → Referral → Revenue.
Map your funnel, find the leakiest stage, and focus experiments there.
Most teams over-invest in acquisition while activation leaks 80% — fix the bucket before pouring more water.
**Viral mechanics.** Viral coefficient (K) = invites per user × conversion rate. K > 1 = exponential growth (rare); K = 0.3–0sustained = meaningful amplifier.
Design for K: make sharing intrinsic to product use (collaboration invites), rewarding (both sides benefit), and frictionless (one-click).
Incentivized virality (pay for invites) usually attracts low-quality users — prefer intrinsic mechanics.
**Onboarding optimization.** Time-to-value is the master metric: measure signup → first key action completion rate and time.
Every field, step, and delay costs activation. Ruthlessly simplify: social signup, progressive profiling, sample data, interactive empty states.
A 10% activation lift compounds through the entire funnel.
## Practical workflow
1. **Diagnose the funnel.** Instrument AARRR. Find the biggest leak relative to benchmarks — that's where experiments go.
2. **Define the North Star.** Pick the metric that best reflects value delivered. Get team agreement; misalignment here wastes quarters.
3. **Build the idea backlog.** Brainstorm across all funnel stages. Score with ICE. Keep 10–20 ideas queued.
4. **Run weekly sprints.** 1–3 experiments per week. Each gets a one-page brief: hypothesis, metric, expected lift, runtime, owner. Ship fast; perfection kills velocity.
5. **Analyze honestly.** Did the primary metric move with statistical confidence? What did we learn? Winners get scaled (roll out to 100%, invest more); losers get killed without sunk-cost drama.
6. **Compound loops.** When an experiment reveals a repeatable mechanism (e.g., invites convert well), productize it into a loop and optimize the loop's conversion at each step.
**Experiment brief template:** hypothesis ("we believe X will cause Y because Z"), primary metric, guardrail metrics, audience/sample, duration, success threshold, owner.
**Growth model (simple version):** New users = acquisition × activation; Retained = new × retention curve; Revenue = retained × monetization. Model each lever's current value and elasticity — invest experiments where elasticity × value is highest.
**Experiment doc template:** hypothesis → metric affected → variant description → audience/traffic → duration → success criteria (pre-registered) → results → learning → next step.
Pre-registration prevents HARKing (hypothesizing after results known) — the silent killer of experiment programs.
**Growth standup (weekly, 30 min):** last week's results → this week's launches → learnings log review → backlog re-prioritization.
Keep it tight; deep dives happen offline. Cadence compounds — never skip.
## Common pitfalls
- **Testing without instrumentation.** Running experiments you can't measure is theater. Analytics first.
- **Acquisition obsession.** Pouring users into a leaky bucket. Fix activation and retention before scaling spend.
- **No statistical discipline.** Calling tests after 2 days, peeking at results, testing 10 variants — all produce false wins.
- **Copying tactics blindly.** "Dropbox did referrals" — your product, audience, and timing differ. Steal the principle, adapt the tactic.
- **Growth at the cost of trust.** Dark patterns and spammy loops boost short-term metrics and destroy long-term brand. Sustainable growth only.
- **Experiment theater.** Running tests to look busy without acting on results. Every experiment must have a pre-committed decision rule.
- **Ignoring retention.** The cheapest growth is keeping users. Churn is a tax on everything else you do.
- **No instrumentation baseline.** Starting experiments without knowing current conversion rates. Measure first, then test.
- **Chasing competitor tactics.** Copying what worked for a different product at a different stage. Principles transfer; tactics rarely do.
- **Growth at all costs.** Optimizing signups while retention collapses. Sustainable growth = retention × acquisition; either alone is vanity.
- **Copying playbooks blindly.** "Dropbox did referrals" — with a product where sharing was intrinsic. Mechanics must fit your product's natural loops.