What is Cohort Analysis?

Cohort analysis groups customers by a shared characteristic (usually signup date) and tracks their behavior over time. It reveals patterns that aggregate metrics hide.

**Example:** Your overall retention might be 85%. But cohort analysis might show:

January cohort: 92% retention

February cohort: 78% retention

March cohort: 89% retention

Suddenly, February requires investigation.

Why Cohort Analysis Matters for SaaS

Problem: Vanishing Metrics

As your customer base grows, new customers dilute your metrics. A company with 80% retention and 20% monthly growth looks healthy — until you realize old customers are churning faster than new ones arrive.

Solution: Cohort Analysis

Cohort analysis lets you:

Compare retention across time periods

Identify the impact of product changes

Measure the quality of different acquisition channels

Predict long-term customer value

How to Build a Cohort Analysis

Step 1: Define Your Cohorts

Most common cohort types:

| Type | Groups By | Use Case |

|------|-----------|----------|

| Time-based | Signup week/month | Standard retention analysis |

| Acquisition-based | Marketing channel | Channel quality comparison |

| Behavioral | First action taken | Onboarding optimization |

| Plan-based | Starting plan | Pricing strategy analysis |

Step 2: Choose Your Metric

Common metrics for cohort analysis:

**Retention rate:** % of cohort still active

**Revenue retention:** % of original revenue retained

**Feature adoption:** % using key features

**Expansion rate:** % that upgraded

Step 3: Set Your Time Period

Track cohorts over:

**Weekly:** For high-volume, short-cycle products

**Monthly:** For most SaaS businesses

**Quarterly:** For enterprise/long sales cycles

Step 4: Build Your Cohort Table

Example Monthly Retention Cohort:

|--------|---------|---------|---------|---------|

| Jan | 100% | 85% | 78% | 74% |

| Feb | 100% | 82% | 71% | 65% |

| Mar | 100% | 88% | 82% | 79% |

| Apr | 100% | 90% | 85% | - |

Interpreting Cohort Data

Pattern 1: Improving Cohorts

Each new cohort retains better than previous ones.

**Indicates:** Product improvements are working

Pattern 2: Declining Cohorts

Newer cohorts retain worse than older ones.

**Indicates:** Quality issues — possibly from new channels or market saturation

Pattern 3: Seasonal Patterns

Certain months consistently perform better/worse.

**Indicates:** Seasonal factors affecting customer quality

Pattern 4: Step Changes

Sudden improvement/decline starting from specific cohort.

**Indicates:** Specific change had impact (price, product, team)

Advanced Cohort Techniques

Revenue Cohort Analysis

Track not just retention but revenue retention:

|--------|---------|---------|---------|---------|

| Jan | $10k | $8.5k | $8.9k | $9.2k |

| Feb | $12k | $9.8k | $8.7k | $8.1k |

Revenue can exceed 100% if expansion outpaces churn (net revenue retention > 100%).

Acquisition Channel Cohorts

Compare retention by how customers found you:

|--------|---------|---------|---------|

| Organic | 90% | 82% | 75% |

| Paid Search | 85% | 71% | 58% |

| Referral | 92% | 88% | 84% |

This reveals which channels bring quality customers.

Using AI for Cohort Analysis

Traditional cohort analysis requires:

SQL queries or spreadsheet wizardry

Manual chart building

Interpretation of results

AI-powered tools like OLARI:

Build cohorts automatically

Identify patterns and anomalies

Explain what's driving differences

Recommend actions based on findings

Getting Started

1. **Export your customer data** — Signup dates and activity

2. **Build a basic time-based cohort** — Monthly retention

3. **Look for patterns** — Any unusual months?

4. **Dig deeper** — Segment by channel, plan, or behavior

5. **Act on insights** — Fix what's broken, double down on what works

Cohort analysis isn't a one-time exercise — it's an ongoing practice that reveals the true health of your business.

Customer Cohort Analysis: How to Track Retention by Signup Date