How do we optimize 'Cohort Analysis' for high-churn subscription models?

All answers to this question.

For SaaS, you should use "Behavioral Cohorts" rather than just time-based ones. Group users by the "First Key Action" they took (e.g., those who finished the tutorial vs. those who didn't). This tells you why they are staying, not just when they joined. Regarding "Reactivated Users," the best practice is to keep them in their "Original Signup Cohort" but track them in a separate "Resurrection" segment. This prevents the "Lazarus Effect" from making a recent month's retention look better than it actually is. Use a "Heatmap" visualization to identify the "cliff"—the specific week where churn usually spikes.

   Answered 2025-11-22 by Margaret Moore

Should we use "Gross Retention" or "Net Retention" when presenting these cohorts to the board to give the most honest view of the business health?

   Answered 2025-11-24 by Brian Allen

• That depends on the story you are telling, Brian. "Gross Retention" shows how well you are keeping the customers you have, which is a measure of product-market fit. "Net Retention" includes upsells and expansions, which shows the "Revenue Growth" within your base. If your Net Retention is over 100%, you have a "Negative Churn" situation, which is the holy grail of SaaS. However, you should always show both. If Gross is low but Net is high, it means you are losing customers but the ones staying are spending more—that’s a risky "leaky bucket" that eventually runs dry.

     Commented 2025-11-26 by Kevin Scott

The "Triangle Chart" is the standard for cohorts. If the colors get darker toward the bottom right, you know your latest product updates are working!

   Answered 2025-11-28 by Nancy King

• Exactly, Nancy. Visual cues are everything in cohort analysis. Thomas, Margaret’s point about "Behavioral Cohorts" is the most actionable advice for a SaaS analyst.

     Commented 2025-11-30 by Thomas Anderson