Multi-Dimensional Cohorts

Single-dimension cohorts reveal trends. Multi-dimensional cohorts reveal causes. If Q4-25 cohort retention is degrading, is it because of enterprise churn, SMB churn, or both? If enterprise retention is stable but SMB retention collapsed, you have a segment-specific problem — not a company-wide one. Multi-dimensional cohorts cross time with segment, channel, product, or deal size to pinpoint exactly where the behavior is changing.

1. Time × Segment Build separate retention matrices for each segment. If enterprise Q4-25 retention at month 6 is 94% but SMB Q4-25 retention at month 6 is 72%, the blended 83% tells you nothing useful. The segment split tells you SMB onboarding or product fit needs immediate attention.
2. Time × Channel Separate cohorts by acquisition source. If partner-sourced customers retain at 91% at month 12 and outbound-sourced customers retain at 76%, the fully loaded cost of outbound acquisition is much higher than the CAC alone suggests — because you are paying to replace a quarter of those customers within a year.
3. Time × Deal Size Segment by initial contract value. Often, the smallest deals have the worst retention and the highest support cost per revenue dollar. A $5K deal that churns after 8 months with $2K in support cost generated $1,333 in gross profit — barely positive. A $50K deal that retains for 36 months generated $90,000. The deal size dimension reveals which customers are worth acquiring.