Cohort Retention Curves and Analysis: Measuring Lifecycle Patterns

Analyzing Cohort Retention Curves

Understanding customer lifecycle patterns. **Building cohort retention analysis** Setup: - Cohort dimension: Acquisition month (most common) - Retention metric: % of cohort still active at each month - Activity definition: Logged in? Used feature? Paid (if SaaS)? - Timeframe: 12+ months (see full lifecycle) Example cohort table: | Cohort | M0 | M1 | M2 | M3 | M4 | M6 | M12 | |---|---|---|---|---|---|---|---| | Jan 2024 | 100% | 95% | 90% | 87% | 85% | 83% | 80% | | Feb 2024 | 100% | 96% | 92% | 90% | 88% | 86% | 82% | | Mar 2024 | 100% | 94% | 89% | 86% | 84% | 82% | - | | Apr 2024 | 100% | 97% | 94% | 91% | 89% | - | - | | May 2024 | 100% | 95% | 91% | 88% | - | - | - | | Jun 2024 | 100% | 96% | 93% | - | - | - | - | Interpretation: - Jan: Steeper decline (5% month 1, then flattens 2-3% monthly) - Jun: Better early retention (4% month 1, similar trajectory) - Trend: Improving retention in recent cohorts Calculating month-over-month retention: - Month 0 to 1: 95/100 = 95% (5% churn) - Month 1 to 2: 90/95 = 94.7% (5.3% churn) - Month 2 to 3: 87/90 = 96.7% (3.3% churn) Churn calculation: - Cohort Jan, monthly churn: - M0-M1: 5% - M1-M2: 5.3% - M2-M3: 3.3% - Trend: Declining churn (good) **Curve shape analysis** Healthy retention curve: ``` 100% ┐ │ ╱╲ (cliff) 90%│ ╱ ╲ │╱ ╲___ 80%│ ╲___ │ ╲___ 70%│ ╲_____ └───────────────────── 0 3 6 9 12 months ``` Characteristics: - Initial cliff (M0-M1): 5-10% churn (normal) - Rapid decline (M1-M3): 3-5% monthly churn - Stabilization (M3+): <2% monthly churn - Plateau (M6+): Flat line (stable, mature customers) Comparison to industry: | Stage | Month 1 | Month 3 | Month 6 | Month 12 | |---|---|---|---|---| | SaaS average | 92-95% | 80-85% | 70-80% | 60-70% | | High retention | 96-98% | 88-92% | 80-85% | 75-80% | | Low retention | 85-90% | 70-75% | 50-60% | 30-40% | Problem curves: Steep cliff (M0-M1): ``` 100% ┐ │ │ ╱╲ 80%│ ╱ ╲___ │ 60%│ ╲_____ └───────────────── 0 3 6 9 12 ``` - Indicates: Onboarding/activation problem - Risk: High early churn (customer not getting value quickly) - Fix: Improve activation, faster time-to-value Continuous decline: ``` 100% ┐ │ │ ╱ 75%│ ╱ ╲ │ ╲ 50%│ ╲___ └───────────────── 0 3 6 9 12 ``` - Indicates: Retention problem throughout lifecycle - Risk: No stable customer base (constant churn) - Fix: Identify what makes customers leave (CS, product) Cohort degradation: ``` Jan Apr Jul 100% │ ━━━ ╱ ╱ 90%│ ╲─╱ ╲───╱ 80%│ ╲╱ └───────────────── 0 3 6 9 12 ``` - Indicates: Recent product change, market shift - Risk: Something broke (product, market, competition) - Fix: Diagnose what changed (features? pricing? market?) **Drilling down on problem areas** Steep cliff (onboarding issue): - Check: Are customers completing activation steps? - Metrics: % completing setup, first feature use, first "aha" moment - Fix: Speed up onboarding, simplify setup, clearer value - Timeline: 2-4 week experiment (test onboarding changes) Month 2-4 churn spike: - Check: When do customers realize value? Why do they leave? - Surveys: Ask churning customers "why did you leave?" - Usage: Are they using core features? - Fix: Improve value realization (CS check-in, more guidance) Gradual decline without stabilization: - Check: Are any customers staying long-term? - Product: Is product solving the problem? - Competition: Any new competitors? Customers switching? - Fix: Feature improvements, CS engagement, pricing alignment Cohort degradation: - Timeline: When did trend reverse? - Changes: What launched/changed around that time? - Examples: - Pricing increase → higher churn - Feature removal → higher churn - New competitor → higher churn - Holiday season → lower engagement - Fix: Revert change? Mitigate competition? Seasonal adjustment? **Segmented cohort analysis** Segment by acquisition channel: | Channel | M1 | M3 | M6 | M12 | Notes | |---|---|---|---|---|---| | Organic | 97% | 88% | 82% | 75% | Best retention | | Sales/AE | 94% | 82% | 70% | 55% | High early value, higher churn | | Paid ads | 90% | 75% | 60% | 40% | Poorest retention | | Partner | 95% | 85% | 78% | 70% | Good retention | Insights: - Organic: More self-selected, better fit - Paid ads: Lower quality leads - Sales: High-touch closes, but not best long-term fit - Action: Shift acquisition to organic/partner (better economics) Segment by customer value (ACV): | ACV | M1 | M3 | M6 | M12 | Notes | |---|---|---|---|---|---| | <£1K | 92% | 78% | 60% | 40% | Lower retention, easier to churn | | £1-5K | 95% | 85% | 75% | 60% | Better retention | | £5K+ | 98% | 92% | 88% | 85% | Highest retention | Insights: - Higher ACV = better retention (more invested) - Action: Focus on higher ACV customers (better LTV) **Predicting and improving retention** Prediction model: - If cohort X has 90% M3 retention (industry avg 80%) - Expect M6 retention: ~82% (apply average decline trajectory) - If M6 actual is 75%, underperforming Leading indicators: - Track M0-M1 churn (early warning signal) - Week 1 activation % (predicts M1 retention) - First 30-day engagement (predicts M3 retention) - CS sentiment (predicts future churn) Monthly monitoring: - Dashboard: Latest cohort M1 retention (is it 95%+?) - Trend: Recent cohorts vs average - Alerts: If M1 churn spikes >8%, investigate - Actions: Monthly retention review, quarterly strategy