Anticipate churn and take action early
With the new Churn Prediction in Customer Intelligence, you can identify customers who are showing signs of churn and understand how likely they are to stop purchasing and when it might happen.
The model uses artificial intelligence to analyze order history and automatically identify your customers' purchasing habits. The prediction is updated every day, following changes in their behavior over time.
This allows you to move from analyzing what has already happened with the RFM model to predicting what might happen next, so you can activate retention and re-engagement strategies in advance.
Six segments to understand who needs action
Based on churn probability and the predicted churn date, each customer is automatically assigned to one of the six churn segments, ordered by increasing level of risk:
- Loyal customers, with a low risk of churn.
- Stable customers, with a moderate risk of churn and no signs of imminent churn.
- Customers to monitor, showing signs that require greater attention.
- At risk customers, with a high probability of churn but still enough time to take action.
- Customers with imminent churn, with a high risk of churn and a predicted churn date approaching soon.
- Churned customers, for customers who have not made any new purchases by the predicted churn date.
This helps you quickly distinguish customers to keep engaged from those who require churn prevention or win-back actions.
Use churn data in your marketing activities
Churn Prediction results are saved in contact fields, so you can use them just like the other data available in magnews.
For example, you can:
- Create audiences of at-risk customers.
- Start workflows when a customer changes segment.
- Personalize communications based on churn propensity.
- Combine churn and RFM clusters to decide which customers to prioritize.
- Use churn probability and predicted churn date to define priorities and timing for retention campaigns.
Monitor the value you can recover
Churn Prediction data is also available on the contact record, in the database dashboard and in audiences.
In addition to the distribution of customers across the six segments, you can monitor indicators such as recovered customers and revenue, customers and revenue at risk, lost revenue and recoverable revenue. This allows you to understand not only how many customers you risk losing, but also the financial impact of churn and the value you can still recover through your retention activities.
Learn how to set up Churn Prediction and read the results to start using the feature in your Customer Intelligence model.
Churn Prediction is available for Customer Intelligence models that use a data table containing individual order history and requires sufficient historical data to generate reliable predictions.