When this skill is activated, always start your first response with the 🧢 emoji.
Growth Hacking
Growth hacking is a discipline that combines product, data, and marketing to find the most efficient levers for sustainable user and revenue growth. Unlike traditional marketing, it is rooted in rapid experimentation, quantitative measurement, and closed-loop feedback between product behavior and acquisition channels.
The best growth practitioners treat retention as the foundation, activation as the multiplier, and virality as the compounding force. Hacks without retention are just churn machines. This skill gives an agent the frameworks, vocabulary, and tactical playbooks to design experiments, build growth systems, and reason about compounding growth.
When to use this skill
Trigger this skill when the user:
- Wants to design or audit a growth loop or viral loop
- Needs to build or improve a referral program
- Asks about optimizing an activation funnel or improving time-to-value
- Wants to reduce churn or improve retention using cohort analysis
- Asks about AARRR metrics, pirate metrics, or north star metric selection
- Needs to run growth experiments and prioritize them (ICE, PIE scoring)
- Is implementing product-led growth (PLG) or a freemium model
- Wants to find the "aha moment" and engineer onboarding toward it
Do NOT trigger this skill for:
- Pure paid advertising campaign execution (creative, ad spend optimization) - use a performance marketing skill instead
- Brand strategy and positioning work disconnected from product or funnel metrics
Key principles
-
Measure everything - Every growth decision must be anchored to data. Define metrics before running experiments. If you can't measure it, you can't improve it. Instrument events, track cohorts, and baseline before changing anything.
-
One metric that matters (OMTM) - Focus each growth phase on a single north star metric that best predicts long-term value. Optimizing many metrics at once diffuses effort and obscures causality.
-
Experiment velocity wins - Teams that run more experiments per week consistently outperform those that run fewer but "bigger" experiments. Lower the cost of an experiment, raise the volume. Most experiments fail - that's fine, fail fast.
-
Retention is the foundation - Acquiring users into a leaky bucket is burning money. Fix retention first. A product with 40% Day-30 retention can grow efficiently; one with 5% cannot be saved by acquisition spend.
-
Sustainable growth over hacks - Short-term hacks (spam, dark patterns, manufactured virality) destroy trust and churn users. Build growth systems that deliver genuine value at each step so growth compounds rather than collapses.
Core concepts
AARRR pirate metrics
Dave McClure's framework maps the full user lifecycle into five measurable stages:
| Stage | Question | Example metric |
|---|---|---|
| Acquisition | How do users find you? | CAC, channel attribution, organic vs paid split |
| Activation | Do users have a great first experience? | Day-1 activation rate, aha moment conversion |
| Retention | Do users come back? | Day-7/30/90 retention, churn rate, DAU/MAU |
| Referral | Do users tell others? | Viral coefficient (K), NPS, referral invite rate |
| Revenue | Do you make money? | MRR, LTV, LTV:CAC ratio, expansion revenue |
Always diagnose which stage is broken before prescribing a fix. See
references/growth-frameworks.md for the full AARRR diagnostic template.
Growth loops vs funnels
A funnel is linear and one-way: Acquire -> Activate -> Retain -> Monetize. Every user enters at the top and exits somewhere below. Funnels are necessary but not sufficient for compounding growth.
A growth loop is circular: the output of one cycle becomes the input of the next. Examples:
- Viral loop: User invites friend -> friend signs up -> friend invites more friends
- Content loop: User creates content -> content ranks in search -> new users find it -> create more content
- Sales-assisted loop: Lead signs up -> sales converts -> expansion revenue funds more sales
Loops compound; funnels don't. Design for loops. See references/growth-frameworks.md
for loop templates.
Viral coefficient (K-factor)
K = invites_sent_per_user * conversion_rate_of_invite
- K > 1: viral growth (each user brings more than one new user)
- K = 0.5-1: strong word of mouth, supplements other channels
- K < 0.3: product is not meaningfully viral; focus elsewhere
Improving K requires either increasing invites sent (motivation) or increasing invite conversion (landing page, offer, trust).
Cohort analysis
Group users by the time period they first performed a key action (signup, first purchase, etc.) and track their behavior over subsequent periods. Cohort analysis isolates the effect of product changes from the noise of a changing user mix.
Key cohort views:
- Retention curve: % of cohort active at Day N - flat curve = good retention
- Revenue cohort: cumulative LTV by cohort - improving means product is getting better
- Activation cohort: % that hit aha moment within Day 1, 3, 7
North star metric
A single metric that best captures the value your product delivers to users AND correlates with long-term business health. It aligns the entire company on what matters.
| Company | North Star Metric |
|---|---|
| Slack | Messages sent per active team |
| Airbnb | Nights booked |
| Spotify | Time spent listening |
| HubSpot | Weekly active teams using 5+ features |
A good north star is: measurable, leads revenue, reflects user value, actionable
by the team. See references/growth-frameworks.md for the selection template.
Common tasks
Design a growth loop
- Map the current user journey end-to-end
- Identify the "output" of one user's experience that could become an "input" for another user (shared content, invites, referrals, SEO-indexed pages)
- Name the loop type: viral, content, paid, sales-assisted, or product-embedded
- Define the loop's single conversion rate to optimize (e.g., invite acceptance rate)
- Instrument every step, establish a baseline, then run experiments on the weakest link
Example - viral loop for a doc tool: Create doc -> Share with external collaborator -> Collaborator views -> Prompted to sign up -> Signs up and creates their own doc -> Loop restarts
Build a referral program
A referral program amplifies natural word-of-mouth with structured incentives.
Design checklist:
- Define the trigger: when is the user most likely to refer? (post-aha moment, post-purchase)
- Choose reward structure: double-sided (sender + receiver both win) outperforms one-sided
- Set reward type: cash, credits, upgrade, or social recognition
- Make sharing frictionless: pre-written message, one-click send, email + link options
- Confirm referral loop is closed: referred user's experience must deliver the same aha moment that motivated the invite
- Track: referral invite rate, referral conversion rate, K-factor, referred-user LTV vs organic LTV
Reward tiers by product type:
- B2C consumer app: credits or cash (Uber, Airbnb model)
- B2B SaaS: seat upgrades, feature unlocks, or billing credits
- Marketplace: transaction credits valid on next purchase
Optimize activation funnel
Activation is the bridge between acquisition and retention. A user is "activated" when they experience the core value of the product for the first time (the aha moment).
Optimization process:
- Define your aha moment concretely (e.g., "creates first project with one collaborator")
- Map every step from signup to aha moment
- Measure drop-off at each step
- Prioritize the step with the largest absolute drop-off (not percentage)
- Run A/B tests: reduce friction (fewer fields, social login), add guidance (tooltips, progress bars), or