North Star Metric (Define First)

Pick ONE metric that:

• Reflects core value delivered to customer
• Leads revenue (not lags)
• Entire team can influence

Examples by business type:

• Marketplace: transactions completed
• SaaS: weekly active users or actions
• Media: time spent or content consumed
• E-commerce: purchase frequency

All other metrics ladder up to this.

AARRR Funnel (Measure Each)

Define specific metrics for each stage:

1. Acquisition: How users find you → visits, signups
2. Activation: First value moment → completed onboarding, first action
3. Retention: Coming back → DAU/MAU, return rate by cohort
4. Revenue: Paying you → conversion rate, ARPU, LTV
5. Referral: Bringing others → viral coefficient, referral rate

Find the weakest stage—that's your focus.

Growth Loops (Build These)

Identify which loop fits your product:

Viral loop: User → invites friends → friends become users

• Measure: viral coefficient (invites × conversion rate)
• Needs: sharing valuable to user, not just company

Content loop: Create content → SEO/social → users → some create content

• Measure: content created per user, traffic per content
• Needs: user-generated content or team-generated

Paid loop: Revenue → reinvest in ads → users → revenue

• Measure: CAC vs LTV, payback period
• Needs: unit economics that work (LTV > 3× CAC)

Sales loop: Sales → customers → case studies/referrals → leads

• Measure: pipeline velocity, referral rate
• Needs: sales team, high ACV

Activation Checklist

Define the "aha moment"—when user gets value:

• What specific action indicates user "got it"?
• How long should it take? (First session? First week?)
• What % of signups reach it currently?
• What steps are required before it?

Remove every obstacle between signup and aha moment. Measure time-to-value and optimize ruthlessly.

Retention Analysis

Cohort retention curves reveal truth:

• Flatten = habit formed, product has value
• Decline to zero = product problem, not growth problem
• Early drop = activation problem

Actions:

• Plot weekly/monthly retention by signup cohort
• Find what retained users did that churned didn't
• Make that action part of onboarding

Channel Selection

Score potential channels:

Prioritize: low CAC + high volume + fast to test first.

Channel categories:

• Paid: Meta, Google, TikTok, influencers
• Organic: SEO, content, social, community
• Product: referral, virality, integrations
• Sales: outbound, partnerships

Test 2-3 max simultaneously. Kill losers fast.

Experiment Framework

For each experiment, document:

• Hypothesis: "If we [change], then [metric] will [impact] because [reason]"
• Metric: specific number you're moving
• Sample size: how many users needed for significance
• Duration: how long to run

Prioritize with ICE:

• Impact (1-10): how much will it move the metric?
• Confidence (1-10): how sure are you it will work?
• Ease (1-10): how fast/cheap to implement?

Run highest ICE scores first.

Quick Wins Checklist

Common high-impact, low-effort fixes:

• Reduce signup form fields to minimum
• Add social proof to landing page
• Implement abandoned cart/onboarding emails
• Add referral program if none exists
• Fix the slowest page load
• Add exit intent offer
• Personalize onboarding by use case

Referral Program Design

Components:

• Incentive: what giver and receiver get
• Mechanic: how sharing works (link, code, invite)
• Trigger: when to prompt (after value, not before)
• Tracking: attribution for rewards

Test: Is the incentive good enough to overcome sharing friction? Double-sided incentives (both get value) outperform one-sided.

Metrics Dashboard

Track weekly at minimum:

• North Star metric
• Funnel conversion by stage
• Retention by weekly cohort
• CAC and LTV (if spending on acquisition)
• Active experiments and results

Segment by: acquisition source, user type, geography.

Common Traps

• Optimizing acquisition when retention is broken—pouring water into leaky bucket
• Too many experiments running—can't tell what worked
• Vanity metrics (signups, pageviews) vs value metrics (activation, revenue)
• Copying competitor tactics without understanding their context
• Not running experiments long enough for statistical significance