# Churn: Defining Churn

This is part 2 of our series on Churn, previous segments are available in our archives.

What is Churn?

Before we can start analyzing our churn, we need to agree on a definition. Churn is typically defined as the rate at which you lose existing customers. In that case, it should be easy to calculate churn in the following way:

However, that isn’t completely clear! When are we measuring customers lost? At what point do we count our total customers? In order to make this work we need to measure churn for a specific time window. As a result, churn is typically measured by month:

That makes more sense, but there is still a big problem. How do we define Total Customers? If we add customers during the month, should we count them? They won’t have had a chance to churn in their first month so that might skew our metric. To address this,churn is calculated only on customers who started the month as customers:

That is better, as now we can understand how many of the customers who started the month were lost during the month. This still won’t capture customers who start and churn in their first month, but at least they won’t skew our calculations.

Even after three iterations this solution has a number of remaining issues:

• Delayed Insight. At any given time the most recent churn metric you will have is at least a month old. In March, the most recent Churn metric you will have is from February! That makes it difficult to address churn problems in real-time.
• Customer Blending. This metric does not discern between customers who have been with you for years and were lost and customers who just recently became customers and left the next month. By blending those values it’s impossible to tell if you are losing long- or short-term customers!

We could continue to refine our equation [1], but, in the end, the churn calculation you use will depend on what you want it to encompass. Do you want churn measurements every week? Do you want to weigh your customers differently? There is no magic, single churn equation because every company considers churn differently.

To create your churn formula, make a list of the important features it needs to encompass and refine it just like we did above. Beware of making it too complex, as the more complex it becomes the more likely someone will make a mistake calculating it at some point and you’ll have a misleading metric.

Tomorrow we’ll cover one way to handle some of these challenges, by combining cohorting and churn.

Do you need help understanding your customers? Outlier is a product designed to help! Outlier looks deep into who your customers are and how they behave to highlight changes that might indicate potential causes of churn. If you’re interested in seeing a demo, schedule a time to talk to us.

[1] If you want to see how far this rabbit hole goes, there is a great post on calculating churn here.

Quote of the Day: “Cultured butter has a slight British accent and impeccable table manners.” – Quora user Eric Mueller on different types of butter