Cohort Analysis: A Complete Guide for SaaS Founders

What is Cohort Analysis?

Cohort analysis groups customers by a shared characteristic (usually signup date) and tracks their behavior over time. It's the best way to understand retention, identify trends, and separate signal from noise.

Why It Matters

Averages Lie

Overall metrics hide important trends. A 5% monthly churn rate could mean:

Steady 5% churn across all cohorts (stable)

Old cohorts at 2%, new cohorts at 10% (problem!)

New cohorts at 2%, old cohorts at 8% (different problem!)

Cohort Analysis Reveals

True retention curves

Impact of product changes

Seasonal patterns

Quality of acquired customers over time

Types of Cohort Analysis

Retention Cohorts

Track what percentage of customers from each cohort remain active over time.

Revenue Cohorts

Track revenue generated by each cohort over time (accounts for expansion).

Behavioral Cohorts

Group by behavior (e.g., feature usage) rather than signup date.

Building a Cohort Chart

Step 1: Define Your Cohorts

Usually monthly signup cohorts:

Jan 2026 signups

Feb 2026 signups

etc.

Step 2: Choose Your Metric

Customer count (retention)

Revenue (revenue retention)

Usage (engagement)

Step 3: Create the Grid

Rows = Cohorts (signup month)

Columns = Months since signup (0, 1, 2, 3...)

Cells = Metric value or percentage

Example Retention Table

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

| Jan | 100% | 85% | 75% | 70% |

| Feb | 100% | 88% | 78% | - |

| Mar | 100% | 82% | - | - |

Interpreting Cohort Data

Healthy Patterns

Retention curves flatten over time

Newer cohorts perform better

Consistent month-over-month retention

Warning Signs

Retention never stabilizes

Newer cohorts worse than older

Steep early drops

Net Revenue Retention (NRR)

Cohort revenue retention can exceed 100% due to expansion:

NRR = (Starting MRR + Expansion - Churn - Contraction) ÷ Starting MRR

Benchmarks

< 80%: Serious retention problem

80-100%: Typical for SMB SaaS

100-120%: Good for mid-market

> 120%: Excellent, common in enterprise

Using Cohort Insights

Product Development

Which cohorts have best retention?

What did they experience differently?

Can you replicate it?

Marketing

Which channels produce best cohorts?

Shift spend to higher-quality sources

Forecasting

Use stabilized retention curves

Project revenue by cohort

More accurate than simple growth rates

Common Mistakes

**Too short timeframes** - Need 6-12+ months of data

**Too small cohorts** - Statistical noise

**Ignoring seasonality** - Compare same months YoY

**Not segmenting** - SMB vs Enterprise may differ significantly

Tools and Implementation

Spreadsheet Method

Works for early-stage, but gets unwieldy.

Analytics Platforms

Amplitude

Mixpanel

Heap

Purpose-Built

Use Valthentic to automatically generate cohort analysis from your data.

Conclusion

Cohort analysis is essential for understanding your business's true health. Start tracking retention cohorts today, and you'll have the insights to make better decisions tomorrow.