Skip to main content

Churn Prediction Invocation

Churn Prediction estimates the likelihood of churn of a subscription within a given horizon. It is based on historical data of subscriptions, and uses machine learning algorithms to provide you with accurate predictions.

Requesting a Churn Prediction Invocation​

To request a churn prediction invocation, you need to have a collection with at least the historical data for subscriptions (but it is highly recommended to provide all the data types). You can create a collection and upload your data following instructions in the Collection section.

Once you have uploaded your data, you can request a churn prediction invocation by sending a POST request to the invocation endpoint, with churn-prediction as invocationType. See more at Invoke Churn Prediction.

Interpreting the Churn Prediction Invocation Results​

The results of a churn prediction invocation request will provide you with the estimated likelihood of customers leaving a not yet canceled subscription.

The outcome of the request to the model will include:

  • id: the identifier of the subscription.
  • churnLikelihood: the likelihood of churn for the subscription.

Here is provided an example for a churn prediction invocation.

Example​

Suppose you have a set of not yet canceled subscriptions (with the IDs: 1, 2, and 3) for which you want to predict the likelihood of churn of the subscriptions.

The output, in json format, will be:

[
  {
    "id": "1",
    "churnLikelihood": 0.187654
  },
  {
    "id": "2",
    "churnLikelihood": 0.875432
  },
  {
    "id": "3",
    "churnLikelihood": 0.543212
  }
]