Updated: May 7
One of the more insightful frameworks for a company to measure user or customer behavior over time is a cohort analysis. In the most basic sense, a cohort is a group, and measuring how it changes over time can produce great insights. It is one of the go-to frameworks we use when evaluating companies at Silicon Road.
A cohort analysis is the measurement of how the behavior of a group changes over time, often month to month. It can be shorter (daily) or longer (quarterly), but we find that a monthly cadence is a good starting point. The framework is used to show trends across metrics like user engagement, active user count and change, revenue, LTV, and churn. It's a window into not just the performance of the company, but also decisions made by the leadership team.
First, let's look at how cohort analysis is presented in figure 1.
Figure 1: Paying customer churn
Customer churn by cohort
This table shows users who started paying in month 1 and how many in that group remained as paying customers in the next month, and the next month, and so on. So, for the cohort of users who started paying in June, 100% of those paying were paying in June. In month 2 (now the month of July), 32% stopped paying, so 68% of the original group remained as paying. By month 6, of that June cohort, just 37% remained paying customers.
So, something we would consider is: how has the business improved retention (reduced churn) of paying customers?
In July, there was some improvement (75% remained paying in month 2), and then something happened in September. Across the board, it appears that something significant happened between August and September to improve retention and reduce churn. This is something that we would want to dig in on as investors. What changed? What led to the decision to change? And so on.
It is a founder's job to represent the business with optimism and belief in the vision of the business. As investors, we study data to provide evidence that this team is making the changes to improve the business. Cohort analysis helps align messaging about the business with the supporting data around that message.
Churn, engagement, customer lifetime value are all important metrics, but without additional context like what cohort analysis framework can provide, these can become more vanity metrics than anything else.
Here's more on cohort analysis and templates to get started: