July 9, 2020

Churning

what does it mean, why does it matter, and how to evaluate it

By
Alex Oppenheimer

Note: this was originally published at https://alexoppenheimer.substack.com/

Anyone who has spent more than 5 minutes thinking or hearing about SaaS companies has certainly bumped into churn. Having spent the last 6+ years neck deep in SaaS companies and their metrics, it is something that I am constantly looking at to assess the health of a business. Three important things to consider when it comes to churn: what does it mean, why does it matter, and how to evaluate it.

What does it mean? In the subscription business model, customers sign up for a contract that involves recurring payments in exchange for a service delivered over time. When a customer ends their subscription, stops paying, and the seller stops delivering service, that is churn. Churn involves a loss of customer and a loss of their associated recurring spend. If a customer stops logging in after pre-paying for a year or goes out of business 6 months into a pre-paid annual contract, it is not churn... yet, but you can bet that it will be when the renewal opportunity comes around.

Why does it matter? The recurring revenue business model is predicated upon recurring customer payments. If customers don't continue paying, then the revenue is not recurring. Every business will likely lose some customers (either voluntarily or involuntarily), but if it's losing a lot of them and it's losing them after a short period of time, it's a big problem as you have lost the power of the recurring revenue business model.

The recurring revenue business model is powerful because it keeps revenue coming in consistently, aligns value delivery with revenue, and makes future revenue streams easy to predict... but only if they actually recur.

If, for example, a company's 2020 revenue target is $10M and they ended 2019 at an ARR of $7M, they only need to "find" $3M of new revenue... assuming they don't lose any of the recurring $7M revenue contracts they ended 2019 with. If they do lose a chunk of it, then it's a larger gap to fill between what they have and what they need to get, increasing variability in the final revenue number.

How to evaluate it. There are three key ways to look at churn and one key way not to look at churn.

Logo or Customer Churn = customers lost in a given period / total customers at the beginning of that period

This tells you whether you lost many customers or just a few, what percentage of your customer base churned, and how does it compare to other periods. Losing many small customers is not great, but might just represent a shift in product or customer segmentation.

Revenue Churn = The ARR dollars lost associated with churned customers in a given period / ARR at the beginning of that period

This tells you whether the customers you lost represent a lot of revenue or a little revenue. Losing one big customer is never good, but is often does not represent systematic problems in a business.

I often look at the average ARR of a churned customer relative to the average ARR over the entire customer base.

Non-Renewals = customers lost in a given period / customers who were up for renewal during that period

This is a critical metric for companies who sell annual contracts, especially if they are growing fast. Notice the vague language I used above about "a given period;" if you are clever about the periods, you can show a nice low churn rate and have a disastrous non-renewal rate reflecting sincere product dissatisfaction and/or a terrible customer base.

As new customer growth decelerates, overall churn rate approaches the non-renewal rate. Reach out to me for a detailed overview of how this happens and what it looks like.

An example: In December 2018 a company adds their first 5 new customers and every month after that they add 5 more customers than they did the previous month. In November 2019 they will have added 60 new customers and have a total of 390 customers at the end of the month. ALL on annual contracts. December 2019 is the first renewal month - what happens? Let's say 4 customers churn. December started with 390 customers, so the churn rate is 4/390 = 1% per month or 12% annualized - pretty good! If you look deeper, you'll notice that there were only 5 total customers who were up for renewal of their annual contracts and FOUR of them churned, which is an 80% non-renewal rate!! While it's great to have kept them for a year, losing 80% after a year is pretty disastrous. Operationally this is how churn actually happens, and therefore a cohort-based approach is absolutely essential for operators to understand what is going to happen in their customer base in the future. Naturally a customer success team should be looking carefully at which customers are up for renewal each month.

How not to look at churn: Net Churn. This is a misleading and useless term in my opinion. Most people calculate it by looking at simple churn % and upsell %, then doing the basic arithmetic to find the difference, and in the process lose a ton of valuable information about the business. In a future post I will talk about cohort-based net revenue retention which IS a valuable metric, but it is calculated very differently and reflects whole cohort behavior over time. It's very important to remember that the most expensive thing for companies is acquiring new customers, so when they lose a customer to churn, no amount of upselling another customer is going to bring back the lost customer, and there is no way to upsell a customer that has churned.

I hope this is a helpful high level overview of churn, how to calculate it, and what it means. I plan on going into much more detail on the calculations, the cohort-analysis, upsell & downsell, and what it means for an operator trying to run their business efficiently and project accurately.

If there are any specific questions or requests, feel free to reach out on twitter @alexoppenheimer


While you're here...
If you enjoyed this post, then you might also like our product — Causal. It's a new way to build models, visualise data, and communicate with numbers. Click here to learn more.