Cohort analysis patterns in Tableau - Step-by-Step Guide

Introduction

Cohort analysis helps you see how groups of customers behave over time. It solves the problem of understanding customer retention and trends by grouping people who share a common event, like their first purchase date.

When you want to track how long customers keep buying after their first purchase

When you need to compare different groups of users who started using your product in different months

When you want to see if a marketing campaign improved customer retention over time

When you want to analyze user behavior by signup week or month

When you want to measure how long it takes for customers to make repeat purchases

Steps

Step 1: Connect

- Tableau Data Source pane

Your data with customer IDs and dates loads into Tableau

💡 Make sure your data includes a date field for the event and a unique customer ID

Step 2: Create a calculated field

- Data pane → right-click → Create Calculated Field

A new field to identify the cohort start date is ready

💡 Name it 'Cohort Month' and use a formula like: DATETRUNC('month', [First Purchase Date])

Step 3: Drag 'Cohort Month' to Columns shelf

- Columns shelf

Cohorts appear as columns representing each group's start month

Step 4: Create another calculated field for 'Months Since Cohort'

- Data pane → right-click → Create Calculated Field

A field showing how many months passed since the cohort start

💡 Use formula: DATEDIFF('month', [Cohort Month], DATETRUNC('month', [Purchase Date]))

Step 5: Drag 'Months Since Cohort' to Rows shelf

- Rows shelf

Rows show time periods since the cohort started

Step 6: Drag a measure like 'Number of Customers' to Text on Marks card

- Marks card → Text

Numbers appear showing how many customers made purchases each month after their cohort start

Step 7: Change Marks type to Heat Map

- Marks card → drop-down → select 'Square' and then Color

Colors show retention intensity, making patterns easy to spot

Before vs After

Before

Table shows raw purchase data with customer IDs and dates mixed together

After

Tableau view shows cohorts by start month on columns, months since cohort on rows, and colored squares representing customer retention counts

Settings Reference

Date truncation level

📍 Calculated field editor

Defines the time unit for cohort grouping

Default: month

Mark type

📍 Marks card

Controls how data points are visually represented

Default: Automatic

Color palette

📍 Marks card → Color

Shows intensity or categories in the heat map

Default: Sequential

Common Mistakes

Using purchase date instead of first purchase date for cohort grouping

This mixes different customers' start times and ruins cohort grouping

Create a calculated field to find each customer's first purchase date and use it for cohort assignment

Not truncating dates to the same level (e.g., mixing days and months)

This causes mismatched cohorts and confusing results

Use DATETRUNC to set all dates to the same level like month or week

Summary

Cohort analysis groups customers by their start date to track behavior over time

Tableau uses calculated fields and date truncation to create cohorts and time intervals

Visualizing cohorts as a heat map helps spot retention trends easily