One of the terms I had never heard before getting into the SaaS business is cohort analysis. As you might expect, a cohort is a set of customers grouped together by common characteristics. The most common types of customer cohorts are by time (e.g. customers that signed up in a given month) and size (e.g. customers based on how much they spend). Cohort analysis is primary used to understand patterns and trends of customer groups over time with the most important metrics being renewal rates and account expansion.
Here are a few thoughts on customer cohort analysis in SaaS:
- Keep cohorts simple while there’s limited data and add complexity as the customer base grows
- Remember that not all customers are equal and the cohorts should reflect a reasonable level of segmentation (e.g. customers by month by size divided into small, medium, and large)
- Consider cohorts on longer tail metrics to see if any insights emerge (e.g. # of logins, module usage, NPS, etc.)
- Look for the “smile” where the revenue expansion of a cohort is expanding (turning up like a smile) vs shrinking (turning down like a frown)
Cohort analysis takes a fair amount of time to initially put together but it’s well worth it — every SaaS company should track their customer cohorts.
What else? What are some more thoughts on customer cohort analysis for SaaS?