All templates
Customer targeting

Churn scoring

Know which customers are ready to churn while there's still time to save them — using Databricks Delta Sharing

You will need a Faraday account to use this template. It is free to sign up and you will just need some sample data to start.

Databricks Delta Sharing logoIf you're a Faraday user already tapping into the power of Databricks Delta Sharing, integrating churn scoring predictions could be a neat advantage. With churn scoring, you identify the customers who might be on the verge of leaving your brand, providing you a chance to re-engage them thoughtfully. In Databricks, the seamless data sharing capabilities mean you can easily incorporate these insights into your data workflows without skipping a beat. You get the insights where you work best and can act on them more efficiently. It's a solid step towards keeping your customer relationships healthy and your efforts well-aligned with your data strategies.
  1. Step 1

    Connect your data sources

    Use the link below to connect Databricks Delta Sharing to Faraday. You can also skip this step and use CSV files to get started instead.
  2. Step 2

    Ingest your data into event streams

    This allows Faraday to understand what your data means. These links will guide you through ingesting the data necessary to power this template.
  3. Step 3

    Organize your customer data

    You'll create groups, called cohorts, that are the essential building blocks of Faraday and allow you to easily predict any customer behavior.
  4. Step 4

    Declare your prediction objectives

    With your cohorts defined, it's easy to instruct Faraday to predict the necessary behaviors. Follow the docs with the link below.
  5. Step 5

    Define your churn scoring pipeline and deploy to Databricks Delta Sharing

    Finally, deploy your prediction with the link below.
  6. Step 6

    Deploy to Databricks Delta Sharing

    Create a deployment target using the Databricks Delta Sharing connection you created above. Or, get started by simply deploying to CSV.