What is Cohort Analysis?
Cohort analysis groups customers by a common starting characteristic (typically the period of acquisition — January cohort, Q1 cohort) and tracks their behavior over time. The classic cohort table shows retention or revenue per cohort across subsequent periods, revealing whether successive cohorts are getting better, worse, or staying flat. Aggregated metrics (overall retention, average ARPU) hide cohort-level trends — a firm whose Q1 cohort retains 70% but Q4 cohort retains 50% may still report flat aggregate retention if Q1 customers are larger. Cohort analysis is foundational to subscription businesses, mobile apps, and SaaS — without it, leadership cannot tell whether the business is improving or coasting on past wins.
How Cohort Analysis actually works
The framework breaks down into the following moving parts. Knowing what each piece is — and what it is not — is what separates a B-grade answer from an A-grade answer in a written assignment.
- Group customers by acquisition period
- Track key metrics (retention, revenue, engagement) over time
- Compare cohorts to identify trends
- Pinpoint period when cohort behavior changed
- Use to evaluate product, pricing, or marketing changes
A worked example: Spotify
Spotify cohort analysis is reportedly one of the most rigorously tracked metrics inside the company. Each monthly cohort's retention curve is monitored — the percentage of users still active at month 1, 3, 6, 12, 24. Comparing cohorts shows whether product changes (e.g., the introduction of personalized playlists, podcast integration, audiobooks) improve retention or not. A new feature that lifts cohort retention at month 6 by even 1% generates millions in CLV across the subscriber base. The cohort discipline is part of why Spotify has consistently grown through competitive pressure from Apple Music and Amazon Music.
Don't lose marks for these
- Reporting only aggregate retention (hides cohort trends)
- Comparing cohorts of different sizes without normalization
- Failing to track long-tail cohorts
How to use this on the exam
Score-maximizing moves
- Show cohort table format
- Distinguish from aggregate metrics
- Apply to subscription business
When to use Cohort Analysis (and when not to)
Use Cohort Analysis when your assignment asks you to analyze, structure, or recommend — and when you have at least two data points to populate every cell of the framework. Skip it when the question is asking for a numerical answer or a single recommendation, since Cohort Analysis is a structuring tool, not a calculator.