Skip to main content




Frequency of Generating Predictions

We generate predictions for what may happen in the next year (one year from when the prediction is generated).

For example, let's say an employee joins in December 2017. If a prediction is generated on a group that includes that employee in June 2018, that employee will be placed in the tenure group that includes employees with 6 months of tenure. If a prediction is generated in December 2018, the employee will be in a tenure group that includes employees with 12 months of tenure. Even though the model has not changed, the prediction for that employee will have changed over the six month period because the attributes of that employee will change from one prediction date to the next.

The predictive models are updated once per year, generally around the end of your organization's financial year. The updated model will likely produce a slightly different prediction than the previous model because more data has become available for the model to learn from.

  • Was this article helpful?