← Back to articles

Cohort Analysis: Tracking Customer Performance by Acquisition Month

Cohort analysis is the most powerful tool for understanding true customer value and revealing hidden trends in your SaaS business. While most companies track aggregate metrics (total churn, total expansion, total revenue), cohort analysis segments customers by acquisition month and tracks how each cohort performs over time. This reveals patterns invisible in aggregated data.

A company's cohort analysis might show that despite stable aggregate churn rate, recent cohorts are churning much faster, indicating product degradation or market shift. Or it might show that expansion is accelerating with recent cohorts, indicating successful upselling improvements. These insights are impossible to detect without cohort analysis.

What Is Cohort Analysis and Why It Matters

Cohort analysis segments customers into groups (cohorts) based on shared characteristics, typically acquisition month. For each cohort, you track metrics (retention, expansion, revenue, unit economics) over time from acquisition.

Example cohort structure: - Cohort Jan 2025: All customers who first became customers in January 2025 - Cohort Feb 2025: All customers who first became customers in February 2025 - Cohort Mar 2025: All customers who first became customers in March 2025 - etc. For each cohort, you track month-0 (acquisition month), month-1 (1 month after acquisition), month-2, month-3, etc. This creates a matrix showing how each cohort evolves over its lifetime.

Why it matters: Cohort analysis reveals whether your business is improving or degrading. Aggregate metrics obscure these trends. A company might have stable 3% monthly churn overall, but cohort analysis might reveal recent cohorts churn at 5% while old cohorts churn at 1%—a clear indication of product/market fit degradation.

Building a Cohort Analysis Table: Step by Step

The most common cohort analysis tracks retention. Here's how to build one: 1. Create a table with acquisition cohorts as rows (Jan 2025, Feb 2025, Mar 2025, etc.) 2. Create columns for month-0, month-1, month-2, month-3, through month-60 (5 years) 3. For each cell, calculate what percentage of that cohort is still active Example table (retention %, 100% = all customers active): Cohort | M0 | M1 | M2 | M3 | M6 | M12 | M24 Jan 24 | 100% | 98% | 95% | 92% | 87% | 75% | 55% Feb 24 | 100% | 97% | 93% | 90% | 83% | 70% | 48% Mar 24 | 100% | 96% | 91% | 87% | 78% | 62% | 35% Apr 24 | 100% | 95% | 88% | 83% | 71% | 55% | — May 24 | 100% | 93% | 85% | 79% | 65% | — | — From this table, you can see: - Month 1 churn is stable (2-5% across all cohorts) - Month 3 churn is accelerating (8-13% across cohorts) - Long-term retention is improving (Jan 24 reaches 55% M24, but Jan 23 might reach only 40%) This reveals actionable insights about onboarding (month 1 is consistent) and product improvements (newer cohorts retaining better long-term).

Revenue Cohort Analysis: Beyond Customer Count

Retention percentage is useful, but revenue-based cohort analysis is more valuable because revenue matters more than customer count. A cohort that retains 50% of customers but has 80% revenue retention (expansion offsetting churn) is healthier than a cohort that retains 60% of customers but has 50% revenue retention (contraction from downgrades).

Build revenue cohort analysis the same way as retention, but instead of tracking "% of customers retained," track "% of starting MRR retained": Cohort | M0 | M1 | M2 | M3 | M6 | M12 | M24 Jan 24 | 100% | 98% | 97% | 96% | 100% | 105% | 110% Feb 24 | 100% | 97% | 95% | 94% | 97% | 102% | — Mar 24 | 100% | 96% | 93% | 91% | 95% | 100% | — This revenue-based analysis reveals very different insights than customer-based: - Customer churn is 3-8% monthly - But revenue is stable or growing (expansion offsetting churn) - This indicates expansion is working well and offsetting customer churn A healthy revenue cohort curve is flat or slowly increasing (customers paying more over time) rather than declining.

MRR Cohort Analysis: The Most Actionable Metric

Beyond percentages, track absolute MRR by cohort to understand revenue contribution: Cohort | Start MRR | M1 | M3 | M6 | M12 Jan 24 | $100K | $98K | $96K | $100K | $105K Feb 24 | $90K | $87K | $84K | $87K | $92K Mar 24 | $110K | $105K | $100K | $105K | $110K This reveals: - Each cohort's absolute contribution to revenue - How much revenue you can expect from cohort 12 months out - Whether you're losing, maintaining, or growing revenue from each cohort To forecast future revenue, sum the M12 MRR from mature cohorts plus estimated MRR for younger cohorts based on historical patterns. This provides more accurate forecasting than top-down growth rate approaches.

Cohort Quality Score: Combining Retention and Expansion

Some companies create a composite cohort quality metric that combines retention and expansion: Cohort Quality = (M12 Revenue Retention % × Customer Retention %) / (M0 Customer Count)

Higher quality cohorts have both good retention and expansion. By month 12, you want to see: - 50%+ customer retention (only 50% of original customers remain) - 100%+ revenue retention (expansion offset/exceeded churn) - Combined quality score reflecting both metrics

Detecting Trends: Product Improvements vs. Market Changes

Cohort analysis reveals whether improvements are working: Scenario 1: Onboarding improvement - M0 and M1 metrics stable across all cohorts - M3+ metrics improving in recent cohorts - Indicates onboarding is consistent but later product experience improved Scenario 2: Product quality degradation - M0-M1 metrics stable - M3+ metrics declining in recent cohorts vs. historical - Indicates product quality issues that don't show up immediately Scenario 3: Market saturation or segment shift - Metric decline across all months for recent cohorts vs. historical - Indicates market conditions changed or you're acquiring lower-quality customers Use these scenarios to guide investigation and action.

Seasonality and Cohort Analysis

Account for seasonal patterns when analyzing cohorts: Budget cycles: B2B customers acquired in January often have high Q1 churn (budget constraints) Holiday effects: Customers acquired December often have lower M1 retention (holiday disruption) Industry cycles: Different for different verticals Compare cohorts to seasonally-similar cohorts (January 2024 vs. January 2025, not January 2025 vs. February 2025). Otherwise, you're confounding seasonality with product changes.

Cohort Analysis by Customer Segment

Run separate cohort analysis for Enterprise vs. SMB customers because they have different retention and expansion patterns: Enterprise cohorts: - Lower M1 churn (customers evaluate carefully before purchase) - Higher expansion (more use cases, departments) - Flatter M12+ curve (mature customers don't change as much) SMB cohorts: - Higher M1 churn (less evaluation, more churn) - Lower expansion (smaller companies, fewer departments) - Steeper decline curve (cost pressure leads to cancellation) Segment-specific cohort analysis reveals which segments are improving with product changes and which are degrading.

CAC and LTV Relationship in Cohort Analysis

Combine cohort analysis with CAC to calculate LTV by cohort: LTV = M24+ Revenue from Cohort × Gross Margin / Number of Customers in Cohort This reveals whether more recent cohorts have lower or higher LTV than historical cohorts: Older cohort (Jan 2024): $120K revenue M24, 100 customers = $1,200 LTV (with 75% margin = $900 profit per customer) Recent cohort (Jan 2025): $150K revenue M12 (projected M24), 120 customers = $1,250 LTV Recent cohort has higher projected LTV, indicating either better product/market fit or lower CAC (or both). Calculate alongside CAC to understand unit economics by cohort.

Communicating Cohort Analysis to Investors

Show investors your cohort analysis with: 1. Retention curves showing M0 through M24+, color-coded by cohort 2. Revenue retention/MRR trends by cohort 3. Trend analysis: "Most recent cohorts show X% improvement in M12 retention" 4. Product improvement narrative: "We implemented feature X in month Y; you can see improvement in cohorts acquired after that date" 5. Segment-specific cohorts if mix is important Cohort analysis demonstrates deep understanding of your business and provides credible evidence of whether product/market fit is improving, stable, or degrading. This is far more compelling than aggregate metrics.

Using Cohort Analysis for Strategic Decisions

Let cohort data drive decisions: Declining M3+ retention across all recent cohorts: Pause customer acquisition growth, focus on product improvement Improving M12+ retention with recent cohorts: Expand customer acquisition, product improvements are working High M1 churn: Improve onboarding, first-time user experience Low M1 churn but high M3 churn: Product disappoints after initial honeymoon period (core feature gaps) Flat revenue curve with high customer retention: Expand pricing/upsell efforts (retention working but expansion failing) Cohort analysis identifies which part of the customer journey is the real problem, enabling targeted improvements.

Tools and Best Practices for Cohort Analysis

Build cohort analysis in a spreadsheet or analytics tool: Excel/Sheets: Create pivot tables with cohorts and months Analytics platforms: Mixpanel, Amplitude, Heap provide built-in cohort analysis Data warehouse: Custom SQL queries on raw customer data Best practices: - Calculate monthly, review quarterly (monthly is noisy, quarterly shows trends) - Account for seasonality when comparing cohorts - Run separate analyses for different customer segments - Pair with CAC and LTV for complete unit economics understanding - Share cohort trends with product and customer success teams to guide their priorities

FAQ: Cohort Analysis

Q: What's the minimum historical data needed to do cohort analysis? A: You need at least 12 months of historical data to see meaningful M12 retention. Ideally 24-36 months to identify trends and seasonal patterns. With less than 12 months, your recent cohorts are incomplete and you can't accurately predict long-term retention.

Q: How do I handle customers acquired through M&A in cohort analysis? A: Exclude them from cohort analysis or run separately. M&A customers have different retention patterns (they already had existing relationship to your company) and inflate numbers artificially. Organic cohorts and acquired cohorts should be analyzed separately.

Q: Should I include or exclude downgraded customers from cohort retention? A: Include them as "retained" but track separately if possible. A customer who downgrades from $200/month to $100/month is retained (still a customer) but contributed less revenue. Retention % includes them, revenue retention % correctly shows the revenue impact.

Q: What if my cohorts are too small to analyze (only 5 customers/month)? A: Aggregate into quarterly or semi-annual cohorts instead of monthly. With only 5 customers/month, monthly cohorts are too noisy. Rolling 3-month cohorts smooth out randomness while still showing trends.

Q: How do I interpret a cohort with 110% revenue retention at month 12? A: Excellent—the cohort is expanding faster than it's churning. Expansion revenue exceeded churn revenue by 10%. This indicates strong expansion motion and product-market fit. This is the goal for healthy SaaS companies.