Growth Hacking

"Growth hacking" is a misleading term — the best growth practitioners are disciplined experimenters, not hackers. The core skill is building a systematic process to identify, prioritize, and run high-velocity experiments across the entire user lifecycle. This isn't a bag of tricks; it's a methodology for compounding small wins into exponential outcomes.

Core Mental Model

The AARRR framework (Pirate Metrics) maps the full user lifecycle. Critically: optimize in reverse order. Improving Retention makes every Acquisition dollar worth more. Improving Revenue means each retained user generates more. Don't pour water into a leaky bucket by optimizing Acquisition before fixing Retention.

AARRR Priority Order (counterintuitive):
5 → Retention     (keeps users you already have)
4 → Revenue       (extracts more value from retained users)
3 → Referral      (turns retained users into acquisition engine)
2 → Activation    (converts signups into retained users)
1 → Acquisition   (only valuable once the above are working)

The leaky bucket metaphor:
Acquisition fills the bucket. Retention is the bucket's integrity.
Fixing Retention multiplies the ROI of every Acquisition dollar.

AARRR Framework In Depth

Acquisition — Where Users Come From

Channels to track:
- Organic Search (content SEO value)
- Paid Search (SEM/Google Ads)
- Paid Social (Facebook, LinkedIn, TikTok Ads)
- Organic Social (viral/referral from social)
- Direct (brand recognition / word of mouth)
- Referral (affiliate, partner, product referral programs)
- Email (outbound, newsletter)
- Product Hunt / Hacker News launches

Key metrics:
- Traffic by channel
- CAC (Customer Acquisition Cost) by channel
- Conversion rate by channel (landing → signup)
- Time-to-first-value by channel (do some channels attract better-fit users?)

Activation — First Value Moment

Activation = user experiences the core value of your product for the first time.
The "aha moment" — the thing that makes them understand why they should keep using it.

Examples:
Slack:     First message sent WITH a coworker (not solo)
Facebook:  7 friends in 10 days
Twitter:   Following 30 accounts in first session
Dropbox:   Saving a file and syncing to second device
HubSpot:   Setting up first contact record
Moltbot:   First message received from another agent

Finding your aha moment:
1. Compare users who retained (D30 > 0) vs those who churned
2. Find the early action that best predicts retention
3. That action IS your aha moment → your activation funnel should drive to it

Activation rate formula:
Activation rate = Users who completed aha moment / Total new signups

Retention — Keeping Users Coming Back

Retention metrics by business type:
- B2C mobile app:  D1/D7/D30 retention
- B2B SaaS:        W1/W4/W12 retention (weekly active accounts)
- Ecommerce:       30/60/90-day repeat purchase rate
- Marketplace:     Supply AND demand retention tracked separately

Healthy benchmarks (vary by category):
Consumer social:  D30 > 25%
Mobile gaming:    D30 > 10%
B2B SaaS:        W12 > 40% accounts

Retention curve interpretation:
Flat line (even at 5%): PMF achieved for that segment
Declining to 0%:        Structural retention problem (wrong audience or product gap)
Rising curve:           Referral or compounding effect (excellent sign)

Revenue — Monetization

Revenue metrics:
- MRR (Monthly Recurring Revenue)
- ARPU (Average Revenue Per User)
- LTV (Lifetime Value) = ARPU × 1/churn_rate
- LTV:CAC ratio target: > 3x (payback period < 12-18 months)
- Expansion MRR: revenue from upsells/seat expansion
- Contraction MRR: revenue lost from downgrades

Growth accounting for revenue:
Net New MRR = New MRR + Expansion MRR - Churn MRR - Contraction MRR

Example:
New MRR:         +$50,000
Expansion:       +$15,000
Churn MRR:       -$20,000
Contraction:     -$5,000
Net New MRR:     +$40,000

Referral — Viral Loops

K-factor (viral coefficient):
K = i × r

i = invitations sent per user (avg)
r = conversion rate of invitations (% who sign up)

K < 1: Product is not self-sustaining via virality alone (needs paid/organic acquisition)
K = 1: For every user, exactly one more user is acquired (linear, not viral)
K > 1: True viral growth (each cohort generates more than previous)

Example:
Average user sends 5 invitations; 20% convert
K = 5 × 0.20 = 1.0 (barely viral — optimize either i or r)

If r improves to 25%:
K = 5 × 0.25 = 1.25 → viral growth loop

Referral mechanic types:
1. Incentivized:     "Give $20, get $20" (Uber, PayPal, Robinhood)
2. Vanity:           "Share your score / ranking" (Spotify Wrapped)
3. Utility:          Product is better when shared (Slack, Dropbox folders)
4. Awareness:        "Sent via [Product]" branding in output (Zoom backgrounds, Typeform)

Channel Experimentation Framework

ICE Scoring for Channels

ICE = Impact × Confidence × Ease (1-10 each)

Channel: TikTok short-form video
Impact:      8 (huge potential reach for B2C)
Confidence:  4 (uncertain — our audience may not be there)
Ease:        3 (requires video production capability we don't have)
ICE Score:   8 × 4 × 3 = 96 (run a small test before investing)

Channel: Google Search Ads
Impact:      7 (reliable, predictable)
Confidence:  9 (we have data from previous campaigns)
Ease:        7 (team knows the tool, high-intent queries clear)
ICE Score:   7 × 9 × 7 = 441 (invest here)

Prioritization: Run ICE on 10+ channel ideas. Top scores get 4-week experiments.

Experiment Design Template

**Experiment Name**: [Short descriptive name]
**Hypothesis**: If we [do X], then [Y metric] will [increase/decrease] by [Z%], 
                because [mechanism].
**Channel/Area**: [Acquisition / Activation / Retention / Revenue / Referral]
**Primary metric**: [Specific, measurable outcome]
**Guardrail metric**: [What we won't let degrade]
**Duration**: [Minimum time to reach significance — calculate, don't guess]
**Sample size needed**: [Based on baseline rate + MDE]
**Effort**: [Low / Medium / High] = [estimated hours]
**Expected result date**: [Date]

Result (fill after):
[ ] Win: Ship to 100%
[ ] Lose: Document learnings
[ ] Inconclusive: Extend or redesign

Learning Velocity

The goal isn't to win every experiment — it's to learn fast. Track:

Experiment velocity: # experiments run per month
Win rate: % that showed statistically significant positive results
Learning quality: Did learnings inform future experiments?

Target for growth-stage company:
4-6 experiments/month per team
25-35% win rate (higher = not ambitious enough; lower = wrong focus)

Activation Optimization

Aha Moment Identification

# Pseudocode for aha moment analysis
# Compare "actions in first 7 days" between retained vs churned users

import pandas as pd

df = events[events['days_since_signup'] <= 7].groupby(['user_id', 'event_name']).size()
df = df.reset_index().pivot(index='user_id', columns='event_name', values=0).fillna(0)

retained = set(users[users['D30_active'] == True]['user_id'])
churned  = set(users[users['D30_active'] == False]['user_id'])

for event in df.columns:
    retention_rate_with    = df[df[event] > 0].index.isin(retained).mean()
    retention_rate_without = df[df[event] == 0].index.isin(retained).mean()
    lift = retention_rate_with / retention_rate_without
    print(f"{event}: {retention_rate_with:.1%} vs {retention_rate_without:.1%} — {lift:.1f}x lift")

# Event with highest "lift" = candidate aha moment

Onboarding Flow Optimization

Principles:
1. TIME TO VALUE: Reduce steps between signup and first aha moment
   - Each step that doesn't drive toward aha = friction = drop-off
   - Show value before asking for information (reverse onboarding)

2. PROGRESSIVE DISCLOSURE: Don't show all features upfront
   - Surface the one feature that delivers the aha moment first
   - Advanced features unlocked/revealed after activation

3. EMPTY STATE DESIGN: A blank product is confusing
   - Seed with sample data, templates, or guided setup
   - "Here's what it looks like when you have data" → skeleton of value

4. COMPLETION ANXIETY: Users don't finish things they start
   - Progress bars increase completion rates (Endowed Progress Effect)
   - Celebrate small wins ("You did it! 1 of 3 setup steps complete ✓")

5. COPY AS UX: The words matter as much as the design
   - Clear verbs: "Add your first team member" not "Team Members"
   - Benefits not features: "Start collaborating" not "Team Settings"

Retention Mechanics

Push Notification Strategy

Rules for push notifications that retain (not annoy):
1. Opt-in moment: ask AFTER the user has experienced value, not on first launch
2. Permission framing: explain the value before the system prompt
   "We'll send you alerts when a teammate mentions you — want them?"
3. Frequency caps: max 1 push/day for most products
4. Personalization: triggered by user's own data, not generic blasts
5. A/B test timing: usually morning (7-9am) or evening (6-8pm) in user's TZ

High-value notification types:
✅ Activity notifications (someone interacted with my content)
✅ Milestone celebrations ("You've used [Product] for 30 days!")
✅ Re-engagement at right moment ("Your project deadline is tomorrow")
❌ Promotional pushes ("Check out our new feature!")
❌ Daily digests nobody asked for

Re-engagement Email Timing

Re-engagement sequence for inactive users:

Day 7 inactive:  "Still getting started?" — tips/resources
Day 14 inactive: "We miss you" — personalized insight from their data
Day 30 inactive: "Your [X] is waiting" — specific hook from their account
Day 60 inactive: Last attempt — offer / strong benefit statement
Day 90+ inactive: Suppress from marketing, exclude from active user counts

Subject line formulas:
Curiosity:  "Something happened to your account while you were away"
Benefit:    "Here's what you've been missing in [Product]"
Social:     "3 of your teammates are using [Product] this week"
Urgency:    "Your free trial ends in 3 days"

Habit Loop Design

Nir Eyal's Hook Model applied to product:

TRIGGER → ACTION → VARIABLE REWARD → INVESTMENT

External trigger: Push notification, email, ads
Internal trigger: Boredom, FOMO, anxiety (strongest)

Action: Tap, open, scroll (must be easy — reduce friction)

Variable reward: 
  - Social rewards: likes, comments, recognition
  - Resource rewards: useful information, deals
  - Self rewards: progress, completion, mastery

Investment: 
  - Storing data, preferences, history
  - Makes product smarter/more personalized over time
  - Switching cost increases with investment

Design for internal triggers:
"When I [feel bored / anxious / want to know X], I open [Product]."
Map your product to a reliable internal trigger.

Conversion Rate Optimization

Landing Page Testing Hierarchy

Test these elements in priority order (highest impact first):

1. Headline — value prop clarity (biggest impact)
2. CTA button — text, color, placement
3. Hero image/video — social proof vs product screenshot vs abstract
4. Pricing display — monthly vs annual toggle, anchor pricing
5. Social proof — logos, testimonials, review count
6. Form length — every additional field reduces conversion
7. Page layout — single-column vs two-column, fold position
8. Copy length — long vs short landing page (depends on product complexity)

Pricing Page Optimization

Pricing page best practices:
- Anchor with highest plan first (rightward anchoring bias → pick middle)
- Highlight recommended plan with "Most Popular" badge
- Annual toggle prominently shown with savings amount ("Save $240/year")
- Feature comparison matrix: include features most buyers care about
- FAQs on pricing page: address top objections (can I cancel? refund policy?)
- Testimonial on pricing page: close to CTA ("We switched from X and save $500/mo")

A/B test ideas:
- 3 plans vs 4 plans (cognitive load)
- Removing free plan (forces upgrade path)
- Feature-named plans vs audience-named plans (Pro vs Teams)
- Monthly as default vs annual as default

Growth Accounting

Monthly growth accounting framework:
Track these 4 user types:

New:         First time active this period
Retained:    Was active last period, active again this period  
Resurrected: Was inactive for 1+ periods, active again this period
Churned:     Was active last period, NOT active this period

Growth rate = (New + Resurrected - Churned) / Prior period active

Quick ratio (product health indicator):
Quick ratio = (New + Resurrected) / Churned

QR > 4:    Exceptional growth
QR 2-4:    Healthy growth
QR 1-2:    Growing but retention problems
QR < 1:    Declining user base

Track both MAU and MRR version of this for full picture.

Product-Led Growth (PLG)

PLG = product is the primary driver of acquisition, activation, and expansion.
      Sales-assist is secondary. Marketing supports awareness.

PLG motions:
1. FREEMIUM: Free tier drives adoption → paid tier for power users
   Examples: Slack, Dropbox, Figma, Notion
   Key: Free tier must deliver real value; paid tier must be compelling enough to convert

2. FREE TRIAL: Full product access for limited time
   Examples: HubSpot, Salesforce
   Key: Enough time to reach aha moment; clear value by day 14

3. REVERSE TRIAL: Free tier users get temporary access to paid features
   Examples: Canva, Grammarly
   Key: Let them experience the premium aha moment; create "loss aversion"

PLG metrics (different from sales-led):
- Time to activate (lower = better)
- Activation rate (% who hit aha moment)
- PQL (Product Qualified Lead) — users who hit expansion triggers
- Product virality (invites sent / accepted per MAU)
- Expansion rate (% of accounts that expand MRR without sales touch)

Anti-Patterns

❌ Optimizing Acquisition before Retention — Pouring more users into a leaky product wastes CAC.

❌ Viral mechanic without product value — "Invite a friend" without a compelling reason to invite creates spam, not growth.

❌ One big bet instead of many small experiments — "The redesign will fix everything" delays learning. 10 small tests beat 1 long shot.

❌ Vanity metric obsession — Page views, app downloads, and social followers feel like growth. LTV-positive retained users ARE growth.

❌ Short experiment windows — Checking results after 2 days and declaring a winner. Run experiments for minimum 2 business cycles (14 days).

❌ Ignoring channel mix — Concentrating 100% of acquisition in one channel creates existential risk (algorithm change, CAC spike, shutdown).

❌ Copying competitor growth tactics out of context — Dropbox's referral worked because their product was worth sharing. The mechanic was downstream of product quality.

Quick Reference

K-Factor Calculator

K = (avg invites sent per user) × (invite → signup conversion rate)

K > 1.0: Viral growth
K = 0.5: Each user brings 0.5 users; need external acquisition to sustain
K = 1.5: Each 100 users brings 150 users; accelerating growth

Experiment Velocity Tracker

Week

Experiments Run

Winners

Learnings

W1	2	0	Headline test inconclusive
W2	3	1	New CTA text +18% clicks
W3	4	2	...

AARRR Quick Diagnosis

Metric down?

Likely cause

First experiment

Acquisition rate	Channel CPM up, landing page CTR down	Landing page headline test
Activation rate	Onboarding friction, wrong users	Reduce onboarding steps
Retention	Wrong aha moment, product gap	Aha moment identification
Revenue	Price too high, wrong tier structure	Pricing page test
Referral	No reason to share, mechanic unclear	Add social proof to share moment