Manu's Vault

Cohort-Analysis

4 min readStatus:Seedling

Cohort Analysis

Cohort analysis is a measurement technique that tracks the behavior of a specific, bounded group of customers — acquired or activated during the same time period — as that group moves through its lifecycle. Rather than averaging behavior across all users at a single moment in time, cohort analysis follows the same customers forward from their entry point.

A cohort is defined by a shared starting event: users who signed up in the same week, customers who made their first purchase in the same month, or users who first logged in during the same product version. Because all members of a cohort entered at the same point, their behavior over time can be directly compared to earlier or later cohorts under different product conditions.

Why Aggregate Metrics Mislead

Without cohort segmentation, product metrics aggregate users at different lifecycle stages. When user acquisition is growing, total engagement will naturally rise — even if per-user engagement is falling. A product can appear healthy in aggregate while each new cohort of customers behaves worse than the last.

Cohort analysis isolates the variable. By comparing the May cohort against the February cohort — measuring the same metric at the same number of days post-signup — you reveal whether product changes actually changed customer behavior. If the April cohort had a 10% activation rate and the July cohort had 12%, the experiments run between April and July were working. If the July cohort has 9%, they weren’t.

This is why cohort analysis serves as the foundation of Innovation-Accounting. Innovation accounting requires a baseline, a series of experiments, and a way to measure whether experiments moved the needle. Cohort comparison provides that measurement mechanism.

Retention Curves

A retention curve plots the percentage of a cohort that remains active at each time interval after acquisition. “Flattening” the curve — where the line stabilizes above zero rather than declining to zero — is a primary signal of product-market fit. Research on consumer apps and SaaS products consistently shows that retained cohorts, not acquisition volume, predict long-term revenue. A leaky bucket filled faster is still a leaky bucket.

What Cohort Analysis Replaces

Cohort analysis replaces Vanity-Metrics-vs-Actionable-Metrics like total registered users, gross revenue, or monthly page views — all of which can improve while the underlying product deteriorates if acquisition is increasing. Actionable metrics must be cohort-aware to be meaningful.

Future Connections

Will connect to Split-Testing, Sticky-Engine-of-Growth when created.

Sources

  • Ries, Eric (2011). The Lean Startup: How Today’s Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses. Crown Publishing. ISBN: 978-0-307-88791-7.

    • Chapter 7 (Measure) — primary source for cohort analysis in the startup context

  • Skok, David (2013). “SaaS Metrics 2.0 — A Guide to Measuring and Improving What Matters.” For Entrepreneurs. Retrieved from https://www.forentrepreneurs.com/saas-metrics-2/

    • Cohort-based churn and LTV analysis; retention curve interpretation for SaaS businesses

  • Chen, Andrew (2018). “New data shows losing 80% of mobile users is normal, and why the best apps do better.” Andreessen Horowitz. Retrieved from https://a16z.com/mobile-retention/

    • Retention benchmarks by app category; statistical basis for interpreting cohort decay curves

  • Fader, Peter S. and Bruce G. S. Hardie (2009). “Probability Models for Customer-Base Analysis.” Journal of Interactive Marketing, Vol. 23, No. 1, pp. 61–69. DOI: 10.1016/j.intmar.2008.11.003

    • Statistical foundations of cohort-based customer lifetime value modelling; formal treatment of acquisition cohort heterogeneity

  • McClure, Dave (2007). “Startup Metrics for Pirates: AARRR!” Master of 500 Hats. Presented at Ignite Seattle.

    • The AARRR framework popularized activation rates as a per-cohort metric — the measure IMVU used in Ries’s original example

Note

This content was drafted with assistance from AI tools for research, organization, and initial content generation. All final content has been reviewed, fact-checked, and edited by the author to ensure accuracy and alignment with the author’s intentions and perspective.

Graph View

Backlinks