How to forecast churn

Losing customers is every business owner's worst nightmare, but understanding the reasons behind it can be the key to unlocking growth and success. This is where churn comes in - a crucial metric that reveals the lost income and customers over a given period.

An important truth for subscription-based businesses is that even seemingly low churn rates can substantially hurt performance.

Churn eats directly into top-line growth and can have severe bottom-line effects.

Customer growth at a 5% growth rate and varying churn rates

As shown above, a 5% monthly churn rate counters the monthly customer growth rate of 5%, leaving the customer count at a stall of 1000.

There are four key questions to answer when forecasting churn:

1. Should you forecast churn based on revenue or customers?

Generally speaking, there are two ways to calculate churn:

You want to use an expected churn rate to estimate how many customers or how much income you expect to churn in each period.

2. How detailed should your churn estimates be?

You need to decide what detail level to forecast churn for.

Examples include

Your chosen method should be based on aknown or presumed pattern.

Example with customer churn

Below are two examples of churn forecasts for a business with three price tiers. The first includes one aggregate churn rate, and the second includes churn rates per price tier.

Churn estimates per price tier may be more appropriate if churn rates differ per tier. Even if you assume the same churn rate across price tiers, you may model out churn per tier to make your model more easily understandable.

Example with revenue churn

Another example is a revenue forecast based on large customer accounts. In this case, it can quickly become arbitrary trying to predict individual customer churn, so it may be more appropriate to forecast churn on an aggregate revenue level.

3. Should you forecast churn on a monthly or annual basis?

Since churn explains lost income/customers over a given period, it’s important to clarify whether it’s monthly or annual.

If customers are locked into a 12-month contract, monthly churn will only show on an annual basis. It might therefore be helpful to split out revenue streams that have different churn patterns.

4. What are your assumed churn rates?

The final step is to predict your specified churn rates. This should be based on historical data or alternatively, an educated guess. 

To get an idea of standard churn levels, this article by Ryan Law (2021) summarizes many public churn benchmarks for SaaS. The article concludes that mature (IPO’d) SaaS companies experience a 5-10% annual churn rate whereas early-stage SaaS experience a 5-10% monthly churn rate. 

At an early stage, uncertain churn assumptions can still provide value as it provides an understanding of what is required to reach financially attractive performance.

Finally - a note on net revenue churn

Net revenue churn is a churn metric that also includes upselling to existing customers, which implicitly makes the above metrics “gross” churn numbers.

Many SaaS companies track net revenue churn to measure and improve their ability to grow customer spending over time. Net revenue churn rates of >100% are ideal.

In conclusion, understanding and forecasting churn is critical to the success of any business, particularly for subscription-based models. Even seemingly low churn rates can substantially impact top-line growth and bottom-line profits. By mastering churn forecasting, businesses can stay ahead of the game, unlocking growth and success.