Likelihood to Churn (LTC) - predicting customer churn

To be released in fall 2021 - set up now to start teaching the algorithm about your metrics.

In order to fight active churn, we want to tackle customer churn before it even happens. In order to do this, ProfitWell has developed a new feature called "Likelihood to Churn" (aka LTC).

Each of your customers will be assigned a LTC score, which will help you understand how likely each customer is to churn in the upcoming calendar month, when compared to all your other existing customers. Think of this as a bell curve like the one displayed below - you can see where a customer falls in the possibility of churning next month. If they are more likely than others to churn (aka more than 1x as likely), then you can proactively do something about it now before they do!

This is a machine learning algorithm, which means that the LTC score will only get smarter in prediction over time as we feed it more of your existing customer activity data. Customer activity data that we use include: total logins, days since last login, total months active, MRR, subscription period length, discounts, geographical regions, and more.

Each prediction is displayed as a probability - we take all the probabilities for each customer and then put them on an index. This is then displayed as a bell curve and the LTC score. The data is retrained each week, and a completely new/updated model is created each calendar month. This means that the LTC scores will update approximately once a week, and the overall prediction each month will get smarter than the last.

Here's how LTC scores look in your dashboard, under each customer's profile:

In order to calculate the LTC scores accurately for your customers, we need at least 3-4 months of Engagement data. This way, we can track the existing activity of your customers. Click here to view your Engagement data or get started.

For additional questions, please feel free to ping us below via chat, or email us at

How did we do?

Powered by HelpDocs (opens in a new tab)