Whenever I talk with Software-as-a-Service (SaaS) entrepreneurs, the topic of growth rates, and corresponding churn rates, always come up (see Quantifying the SaaS Valuation Growth Rate Multiplier). After talking about churn rates, I like to ask the following question: how long do customers have to stay with you to know they’re in it for the long haul?
For the first few years of Pardot, most customers were month-to-month without an annual contract. At first, we paid commissions out to our sales reps at a rate of roughly 12% of the first year’s revenue. Quickly, we realized that if customers weren’t a good fit, they’d leave within the first four months. Customers that stayed longer than four months would be customers indefinitely. To make things more aligned, we changed our commission with sales reps to be 50% of the first four months of customer revenue, paid monthly as revenue came in, so that the sales reps would sign good-fit customers.
Now, there’s a debate in the SaaS metrics world about calculating churn. Do you count all new customers signed or do you only count new customers that have made it past the critical starting period (e.g. the first four months in the early Pardot example)? As a course of business, it’s easier to count all new customers as that makes it easy to track and measure the critical SaaS metrics. For the health and long-term view into the business, I prefer not counting customers that churn in the critical starting period. Cost of customer acquisition, renewal/churn rates, lifetime value of the customer, etc all have to take into account not counting early churn customers, if that’s the route the entrepreneur chooses to take regarding metrics.
Some variances of this practice include analyzing cohorts on a weekly/monthly/quarterly basis as well as only counting customers if they meet certain size requirements (e.g. it’s usually the smallest of businesses that have the highest churn rates, so if you exclude any customers below a designated size when tallying the metrics, they don’t negatively affects the startup’s real numbers). Put another way, don’t assume that all new customers should be tracked the same way from a metrics perspective.
What else? What are some more thoughts on measuring customers that churn early as part of the overall customer metrics?