Cohort Heatmaps Basics

Why this matters

Cohort heatmaps help Marketing Analysts see how groups of users behave over time after a shared starting point (like signup or first purchase). This is essential to answer questions such as:

• Are users retained after their first week or month?
• Do customers who joined after a campaign repurchase at higher rates?
• Which onboarding email changes improved week 2 engagement?
• How does a new pricing plan affect month 3 churn?

Concept explained simply

A cohort is a group of users who share a start event and date (e.g., first purchase in March). A cohort heatmap is a grid: rows are cohorts (by start period), columns are periods since that start (Week 1, Week 2… or Month 1, Month 2…), and each cell shows a metric (like retention %) colored by intensity.

How to build one (step-by-step)

1. Choose your cohort key: Typical options: signup month, first purchase month, acquisition campaign start.
2. Pick the time grain for columns: Weeks since start or months since start. Keep it consistent.
3. Define the metric clearly:
   • Retention rate: active users in period k / total users in cohort.
   • Repeat purchase rate: customers with ≥1 order in period k / total customers in cohort.
   • Revenue per user: total revenue in period k / total users in cohort.
4. Transform your data: For each user, tag their cohort and the period number (k) since cohort start; aggregate your metric by (cohort, k).
5. Pivot to a matrix: Rows = cohorts, Columns = period k. Fill cells with your metric.
6. Color-scale with care: Use a sequential scale (light to dark) for 0% to 100% metrics; add value labels or tooltips where possible.
7. Handle empties vs zeros: Show missing data (not enough time elapsed) as blanks, not 0.

How to read a cohort heatmap

• Scan diagonally down-right to see how a cohort decays or grows over time.
• Scan vertically in a column to compare different cohorts at the same age (e.g., Month 2 retention across cohorts).
• Look for structural changes: a brighter column after a product update implies improvement at that age for newer cohorts.

Worked examples

Common mistakes and self-check

• Wrong denominator: Dividing by active users in prior period instead of the full cohort. Self-check: does Month 0 equal 100%? Later months should be ≤100% for rates.
• Calendar vs relative time: Using calendar months across columns. Self-check: columns must be “months since start.”
• Mixing cohort keys: Some rows by signup, others by first purchase. Self-check: confirm one consistent start event.
• Color scale misleads: Diverging red–green for unipolar data. Self-check: use a single-hue scale and show a legend.
• Zeros vs missing: Empty future periods shown as 0. Self-check: blanks or dashes for not-yet-available data.
• Small cohort noise: Reading too much into tiny cohorts. Self-check: annotate cohort sizes; consider smoothing or minimum size filters.

Exercises you can do now

Do the exercises below and use the checklist to verify your work.

• Exercise 1 (ID: ex1): Compute weekly retention for three cohorts from a small user log.
• Exercise 2 (ID: ex2): Write correct formulas for repeat purchase rate and revenue per user.

Practical projects

• Build a monthly retention heatmap for website signups from the last 12 months. Include cohort sizes in the row labels.
• Create a repeat purchase heatmap for first purchase cohorts using Month 1–6. Add a separate row note for major campaigns launched.
• Make an engagement heatmap for trial users: Week 1–8 active sessions per user. Cap outliers and annotate thresholds.

Who this is for

• Marketing Analysts who need to track retention, repeat purchase, or activation quality.
• Growth and Lifecycle marketers who own onboarding and CRM performance.
• Product-leaning analysts visualizing behavior after feature launches.

Prerequisites

• Basic data manipulation skills (grouping, counting, aggregating).
• Comfort with percentages and ratios.
• Familiarity with pivoting data into wide format.

Learning path

1. Understand cohorts and time alignment (this lesson).
2. Calculate retention and repeat purchase metrics.
3. Create a clean matrix and apply readable color scales.
4. Interpret diagonals and columns to generate insights.
5. Communicate findings with short, actionable notes.

Next steps

• Complete the exercises below and take the Quick Test.
• Apply the method to one real dataset at work and share a 3-bullet insight summary.
• Iterate with a second metric (e.g., revenue per user) to deepen the story.

Mini challenge

Pick a dataset you know. Define your cohort key, time grain, and one metric. Draft a 4-row by 4-column cohort table and write two observations: one about a diagonal trend, one about a specific column comparing cohorts.

Note: The Quick Test is available to everyone; only logged-in users will have their progress saved.