Help Center

Analyze customer churn propensity with AI Workbench

How to analyze customer churn with AI and machine learning models in the BlueConic AI WorkbenchBlueConic provides several prebuilt AI notebooks you can use (without writing any code) to apply the power of AI to your customer order data, for example to predict and analyze a customer's churn propensity. 

AI Workbench provides two notebooks relating to customer churn predictions: the Predict propensity to churn notebook to calculate customer churn predictions, and this notebook to compare churn rates for different segments of customers. Both notebooks assume you already have customer contract or subscription start and end dates data present in separate BlueConic profile properties.

Adding a churn modeling notebook in BlueConic

  1. Select AI Workbench from the BlueConic navigation bar.
  2. Click Add notebook.
  3. A pop-up window appears. Scroll down to the Analyze propensity to churn notebook and click it.
    How do I analyze customer churn with a CDP and AI Workbench?
  4. The notebook opens to the Parameters window.
    Note: If you have the Notebook editor permissions, the Notebook editor window is where you can view the notebook's Python code. It also contains detailed documentation about how the notebook works, and how the machine learning model uses customer transaction data for running and analyzing customer churn predictions. 

Setting parameters for the Analyze customer churn propensity notebook

  1. Enter a name for the notebook and click Save.
  2. Select the Parameters tab in the left-hand panel of the churn prediction notebook window.
    How to run AI models for customer churn in BlueConic
  3. For the segments to compare parameter, choose one or more customer segments for the notebook to predict churn probability.
    The example shown here compares the churn risk for three customer segments: all customers, those with broadband contracts, and those with mobile phone contracts.
    How do I use BlueConic CDP to analyze customer churn with AI Workbench?
  4. Select profile properties that contain these profiles' start date contract and end data contract. These profile properties should be of date time format.
  5. (Optional) You can adjust the maximum number of profiles the model will examine. This defaults to 2500 profiles.
  6. (Optional) You can change the propensity to churn model used to display your results. See below to learn more about each of the propensity to churn models available for visualizing your results.
  7. Save your settings before running the model.

Selecting a propensity to churn model

By default, the AI notebook uses the Kaplan-Meier estimator to display your results. In the Parameters tab, you can choose from among several different models (and color schemes) to visualize your results. 

Running an Analyze propensity to churn notebook

  1. Go to the Schedule and run history page.
  2. In the metadata section at the top of the page, you can request email notifications each time the notebook runs or only for failed runs. For details, see: setting up email notifications for AI Workbench.
  3. Click Run now to run the analysis manually.
  4. To schedule the import and export for a future date, activate Enable scheduling. Click the Settings icon mceclip4.png. Select how to schedule the import by choosing an option from the dropdown list:
    • Every X minutes
    • Number of times per day
    • Days of the week
    • Days of the month
    • Weekday of the month

    Set a time for the import. Click OK


After running the notebook, you can view its output by clicking Preview. Scroll all the way to the bottom to see a graphical representation of the churn analysis. You can use the Notebook all cells insight to create a sharable dashboard with churn prediction graphics.

Companion AI notebook for predicting customer churn

You can use the companion AI notebook, Predict propensity to churn, to examine how churn risk compares among groups of customers within a single customer segment. The notebook also writes a churn prediction score to a profile property.

Was this article helpful?
0 out of 0 found this helpful