Cohort Analysis Guide

Quick Start Guide for Cohort Analysis.

Cohort Analysis Workflows

Example: Monthly Customer Retention

Scenario: You're a SaaS company and want to understand how many customers remain active over time, grouped by the month they signed up.

Your Data

You have transaction data in "Periods as Rows" format:

customer_idmonthrevenue
alice@co.com2024-01100
alice@co.com2024-02100
alice@co.com2024-03100
bob@co.com2024-0150
bob@co.com2024-0250
carol@co.com2024-02200
carol@co.com2024-03200
carol@co.com2024-04200
dave@co.com2024-0375

Step-by-Step Setup

  1. Open Jetti and select your data sheet
  2. Column mapping:
    • Customer ID → customer_id
    • Date/Period → month
    • Value → revenue
  3. Settings:
    • Report type: Cohort
    • Value type: Count (we want to count customers, not sum revenue)
    • Interval: Monthly
  4. Generate Report

Expected Output

CohortMonth 0Month 1Month 2Month 3
2024-012210
2024-02111
2024-0311

As percentages:

CohortMonth 0Month 1Month 2Month 3
2024-01100%100%50%0%
2024-02100%100%100%
2024-03100%100%

How to Interpret

  • 2024-01 cohort: Alice and Bob both signed up in January. Both were active in February (100% retention). Only Alice remained in March (50%). Neither was active in April (0%).
  • 2024-02 cohort: Carol signed up in February and stayed active through April (100% retention at Month 2).
  • 2024-03 cohort: Dave signed up in March and was still active in April.

Key Insights

  • The January cohort shows 50% drop-off at Month 2 - what happened to Bob?
  • February and March cohorts show better early retention
  • Click on the "50%" cell to drill down and see it was Bob who churned

Example: Revenue Retention by Cohort

Scenario: Same data, but now you want to track how much revenue each cohort retains over time (not just customer count).

Setup Changes

  1. Keep all settings the same, but change:
    • Value type: Sum (sum revenue instead of count customers)

Expected Output

CohortMonth 0Month 1Month 2Month 3
2024-01$150$150$100$0
2024-02$200$200$200
2024-03$75$75

As percentages of starting revenue:

CohortMonth 0Month 1Month 2Month 3
2024-01100%100%67%0%
2024-02100%100%100%
2024-03100%100%

How to Interpret

  • January cohort started with $150 (Alice: $100, Bob: $50)
  • By Month 2, only $100 remained (Alice) - 67% revenue retention
  • The 2024-02 cohort (Carol at $200) retained 100% through Month 2

Revenue vs Count: When to Use Each

Use Count when...Use Sum when...
You care about logo retentionYou care about dollar retention
All customers are roughly equal valueCustomers have varying contract sizes
You're tracking user engagementYou're tracking revenue/MRR

How Cohort Calculations Work

Step 1: Determine Each Customer's Cohort

Jetti assigns each customer to their first active period:

customer_idFirst seenCohort
alice@co.com2024-012024-01
bob@co.com2024-012024-01
carol@co.com2024-022024-02
dave@co.com2024-032024-03

Step 2: Calculate Cohort Age

For each period a customer is active, calculate how many periods since their cohort:

customer_idCohortActive PeriodCohort Age
alice@co.com2024-012024-010
alice@co.com2024-012024-021
alice@co.com2024-012024-032
bob@co.com2024-012024-010
bob@co.com2024-012024-021
carol@co.com2024-022024-020
............

Step 3: Aggregate by Cohort and Age

Jetti groups customers by their cohort and age, then counts (or sums) the values for each cell in the table.

Step 4: Calculate Retention Percentage

Divide each cell by the Month 0 value for that cohort:

Retention = (Value at Month N) / (Value at Month 0) × 100