Last week, Jason Cohen wrote a very comprehensive blog on software-as-a-service churn: Deep Dive – Cancellation Rate in SaaS Business Models. I required everybody at Totango to read this blog and recommend that you do the same. Jason looks at many different definitions for the SaaS Cancellation Rate metric.
Eventually, Jason recommends performing cohort analysis when looking at cancellation rates. He suggests to divide customers in segments based on their “time to cancel” (i.e. cancelled after 30 days vs. cancelled after more than 30 days) and, for all intends and purposes, he recommends focusing in the long-term users who have greater business revenue potential and cancellation reasons which can be addressed and resolved more easily.
This is indeed an interesting way to look at it, and very analogous to the importance of the “time to convert” metric when it comes to inbound marketing and trial conversion. However, I argue that this is not the only, and maybe not always the best, way to do cohort analysis on SaaS churn.
Let’s take for example an email service application. If 2 users have signed up at the same time:
- One of them is using the service more frequently, creating many accounts, visits almost all application features and cancels after 10 days
- The other accesses the service 3 times a week but just checking very limited features and cancels after 31 days
Who should be given more weight?
If I’d measure by Jason, I would focus my efforts on the second user, but if I weigh my analysis with user behavior altogether, then my most valuable customer to understand is the first one.
So this leaves us with three promising ways to segment customers for cohort analysis:
- Traditional way: create cohorts based on the week or month in which they signed up for the service. This will allow you to analyze the effect of changes you made to your product or service over time.
- Jason’s way: to create cohorts based on the “time to cancel” (or the “time to convert” for that matter). This will allow you to focus on long-time users of your product and sift out those who signed up in error.
- The customer engagement way: to create cohorts based on the “engagement level” with the product or service. This will allow you to focus on frequent users of your products, independent on how long it took them to cancel, but still sift out those who signed up in error (and never started to use the product).
Of course, in all cases, measurement is just the first phase of the process and the complementary phase must be to prioritize the changes needed in the service which would ultimately lead to increase customer satisfaction and customer engagement.
What about you? What is your definition for cohort analysis?