AI summary by TwinMind
Consumer Startup Metrics | Startup School
Y Combinator
22m 26s
Growth Metrics Overview
- Good growth rate: 15% month-over-month (5x user base annually)
- Acceptable growth: 10% monthly (triple user base annually)
- Poor growth: 5% monthly or lower (unlikely to reach breakout success)
- Growth complexity extends beyond headline user numbers
Organic vs Paid Growth Split
- Organic Growth Definition: Anything not paid for directly
- Monzo Example: Reached 1 million customers before spending on direct marketing/advertising
- Best Companies Split: 80%+ organic to 20% paid, even 100% organic (Facebook, WhatsApp early days)
- Acceptable Split: 50/50 organic/paid is okay
- Warning: Below 50% organic for extended periods is worrisome
Virality Mechanisms
- Definition: One user using product introduces it to other users through product use
- Facebook Example: Photo tagging friends not on platform → email notification → viral signup
- Wordle Example: Sharing score grids on social networks attracts new players
- Key Questions:
- What are sharable moments in your product?
- How to make sharing easy with native sharing prompts?
Network Effects
- Metcalfe's Law: Network value = square of nodes in network
- WhatsApp Example: More users = more messaging value
- Monzo Example:
- Venmo-style money transmission within bank
- Joint accounts and group "pots" for holidays
- Groups of 6-7 people: 3 have Monzo → all get "bullied" to sign up
- Key Shift: From single-player to multiplayer product journey
Viral Loops vs Ad Spend ROI
- Viral loops and network effects pay back forever
- Ad spend: Money spent today, gone tomorrow
- Every 1-2% optimization in viral loops pays back for company lifetime
- Focus on optimizing these mechanisms over paid acquisition
Paid Referral Schemes
- Examples: Member-get-member ($5 each), Uber free rides
- Treatment: Consider as paid acquisition, not organic
- Cannibalization Risk: Paying for users who would have referred anyway
- Testing Method: Run schemes in specific cities/regions or turn off periodically
- Fraud Risk: People gaming referral systems (Zipcar Google ads example)
Paid Growth Tracking
- Essential: Know where every user comes from
- Methods: UTM referrals in URLs, ask users "where did you hear about us?"
- Avoid: Overly complex attribution models (first-touch, last-touch, multi-touch)
- Database Recording: Store acquisition source permanently for each user
Customer Acquisition Cost (CAC)
- Definition: Cost per channel to acquire each user
- Track Per Channel: Monitor performance over time
- Monzo Money-Saving Blog Example:
- Cheap acquisition cost initially
- Users loaded money, went on holiday, withdrew £1000 cash immediately
- Negative lifetime value → shut down channel
- Measure to Active Users: Track cost to generate retaining, monetized users, not just signups
- Account for Drop-off: 80-90% first week drop-off rates common
Paid Growth Risks
- Platform Dependency: Google and Meta extract maximum value from advertisers
- Bidding Wars: Competitors bid up costs, margins go to zero
- Platform Changes: iOS tracking changes can wipe out half the business overnight
- Scale Limitation: No great consumer company reached IPO scale with >50% paid signups
Unit Economics Framework
- Definition: Per-customer revenue minus variable/incremental costs
- Monzo Example:
- ~$50 revenue per customer
- Variable costs: plastic card shipping, customer service, fraud, transaction fees
- Importance: Understand profitability per customer and cohort
- Optimization: Invest in channels bringing profitable customers, reduce costs
Variable vs Fixed Costs
- Variable Costs: Scale with customer count (cards, support, fraud)
- Fixed Costs: Don't change with customer doubling/tripling (engineering salaries, office rent)
- Unit Economics: Revenue minus variable costs only
- Monzo Scale: -£3-4 per customer per year, scaled to 500k customers before fixing
Consumer Retention Challenges
- B2B vs Consumer: B2B has clear churn signals (contract cancellation)
- Consumer Complexity: Periodic usage patterns (Airbnb vs Facebook)
- Usage Frequency Examples:
- Facebook: Multiple times daily
- Airbnb: Every 6 months
- Monzo: 1 transaction per week minimum
Magic Moment Identification
- Definition: User behavior predicting long-term retention
- Analysis Method: Compare best users' behavior to rest of customer base
- Facebook Example: 7 friends added in first 10 days → long-term retention
- Monzo Example: 3 friends added from phone book → 20 percentage points better retention
- Product Re-engineering: Optimize onboarding to hit magic moment
- Avoid Over-optimization: Don't fixate on exact numbers (7 vs 6 vs 8 friends)
Net Promoter Score (NPS)
- Scale: 0-10 likelihood to recommend
- Calculation: % Promoters (9-10) minus % Detractors (0-6), ignore neutrals (7-8)
- Score Examples:
- All detractors: -100
- All promoters: +100
- 50% promoters, 25% detractors: +25
- Minimum Baseline: +50 for new consumer companies
- Monzo Range: +75 to +80
- Tesla Current: +96
- Incumbents: Often 0 or negative (telecom, old banks)
NPS Collection Best Practices
- Consistency Critical: Same collection method, timing, and customer lifecycle point
- Grouper Example: Changed collection method → 20 point overnight decrease
- Follow-up Question: "Why?" for qualitative feedback
- Improvement Strategy: Fix detractor complaints to increase score
- Frequency: Monthly random samples or specific product moments
Key Benchmarks Recap
- Growth Rate: 15% monthly minimum, 10% acceptable, 5% insufficient
- Organic/Paid Split: 50%+ organic required for scale
- Unit Economics: Must be positive before scaling
- Retention: Must flatten somewhere, use magic moments for long cycles
- NPS: +50 minimum for great consumer startups
Final Warnings
- Industry Variation: Benchmarks may differ by company/industry
- Trade-offs: May have great organic growth but poor unit economics (early Monzo)
- Business Uniqueness: Consider how metrics apply differently to your specific business
