Saas Math: Customer vs. Revenue Churn

As I’ve mentioned many times before: churn (the rate at which customers cancel their subscription) is the most important metric for any recurring revenue business. This should make sense. The longer a customer keeps paying you, the more valuable that customer is.

In a previous post, we looked in depth at churn. Most companies when they look at churn look at customer counts: how many customers did we start the month with? How many of those did we lose? Divide the lost customers by the beginning customers and you have a basic measure of churn.

While it’s important to understand this and to look at these patterns over time on a cohort basis (i.e. for customers who joined in a given month, from a common source, or linking them via any other common attribute to find patterns), you need to also look at revenue churn.

Customer churn then is a count of customers who cancel their accounts. Revenue churn is revenue you lost from those customers.

Here is a simple example:

This example looks at changes in customer count per month. It then breaks out Monthly Recurring Revenue (MRR) based on:

– How much MRR the company had at the beginning of each month;

– How much new MRR it gained (from converting new customers and from upgrades by existing customers); and

– How much MRR it lost from customers that canceled their accounts.

From these stats, we can calculate both customer and revenue churn.

What you should always see is that customer churn is > revenue churn.

Why? Well, most (all?) SaaS companies offer a variety of price plans for their offerings. Everyone seems to have that 37 Signals‘ style price plan with 5 options where the middle one is magically the most popular (this is almost always not true). Sorry, I digress…

As I was saying, if you have multiple price points for your offering, chances are very good that most of the customers you are losing are on the lowest price point. This is because most churn occurs in the first 30 – 60 days. And these newer customers tend to be on your lowest price plan. So, even though they count the same as any other customer in terms of customer churn, they are worth a lot less than your customers that have been with you for a while.

When you break down your revenue churn you should see much lower churn on your higher price plans and in terms of customer tenure you should be seeing much higher churn for newer customers. If these patterns are not evident every month, then there is either an issue with your data or your product.

Churn by Price Plan

You can take your analysis further and look at churn per price plan. When you do this, you might come to some startling conclusions.

For example, when you look at per user economics by price plan (looking at revenue from users on each price plan less hosting & service costs less customer acquisition costs), you might find that your entry level pricing plan does not give you profitable customers. Yes, it adds to customer count and revenues, but these customers might not be contributing to your bottom line. These customers also tend to chew up more support, further reducing their profitability.

Conversely, higher price point customers may be so valuable (because they are more established and stick around more) + lower maintenance that you decide to launch a program to get customers to upgrade or go after higher value customers.

These are patterns I have seen before, but they not apply to your business.

The point is: slice and dice your data every way possible to get the insights you need to optimize your business.

  • Hi Mark,

    Insightful read, thanks. I have a question related to the relationship between these 2 metrics. If you were to put ONLY one of these metrics in a dashboard (which will contain other key metrics so there's pressure to display only the critical few to keep it manageable) which one would it be? Customer churn or revenue churn?

    Looking at it from a business point of view, my pain point is lost revenue of which customer churn is a driving factor (but not as painful). If I didn't lose revenue (or did only by a fraction), I wouldn't lose sleep over customer churn.

    So in that sense, would you say that the more critical metric to monitor is "revenue churn" while "customer churn" would be better used when analysing the evolution of "revenue churn"? So, rather than put them both on the dashboard, does it not make more sense to:

    1. first establish which is the more relevant of the two metrics
    2. establish the cause-effect relationship between the two metrics
    3. break the performance of the critical metric down by the less critical one in order to understand why it's gone up, down or stayed the same?

    In other words, which of these 2 metrics, at a glance, tell you that the business is in trouble?

    • Revenue churn is the most important one. Knowing the long term value of a user helps you figure out how aggressive to be in customer acquisition cost. Customer churn doesn't give you the right data for that

  • Pingback: This Week in Canadian Startups - Mark Evans TechMark Evans Tech()

  • Pingback: SaaS - Customer acquisition costs | Pearltrees()

  • Great read. Segmenting users by price plan and other demographic traits is really powerful. It's also fascinating to flip the tables and look at behavioral segments. For example, if your goal is to reduce churn, then segment all the users who've churned and analyze them: referrer, usage behavior, errors encountered, etc. If you compare this to long-time users, you'll see immediately how to reduce churn.

  • Nice post Mark.

    Not to sound like an total product plug but we use KissMetrics on our product and it does an amazing job of providing insight into both customer and revenue churn ( It's incredibly understanding the differences in churn and LTV among different plans, billing amounts, and referral sources.

    • Hey Tyler,

      Yep, many of my companies use Kiss. Works well. Like your product BTW. Feel like the pricing could be higher. Hard to do much paid customer acquisition at that monthly price point.

  • Kyle S

    Hi Mark –
    I enjoyed this post but am having trouble with your math. In month 1 it looks like there should be a net gain of 30 customers (150 adds less 120 cancels) yet that is not what the math shows. What am I not getting?

    • Hey Kyle. It was a cut and paste error. I had changed churn to better illustrate the difference between customer vs revenue churn but did not change the formulas all the way through. thanks