The page provides the default definitions of Predictable’s scoring populations for each model.

Summary Table

MODEL TYPE HISTORICAL WINDOW PREDICTION WINDOW ADDITION CRITERIA
PURCHASE PROPENSITY 90 30 Must have made a purchase / clicked email / triggered web event within historical window
CHURN 365 180 Must have made a purchase within historical window
LTV 365 365 Must have made a purchase / clicked email / triggered web event within historical window
NEXT PURCHASE 180 90 Must have made first purchase within last 14 days
PRODUCT RECOMMENDATION 365 N/A Must have made two purchases within historical window

Purchase Propensity

Scoring Criteria

In order to be scored by the Purchase Propensity model, a customer must have done one of the following actions within the last 90 days:

  • Made a purchase
  • Clicked an email
  • Triggered a web event

If a customer does not fit this criteria, they are dropped out of the scoreable population and will not receive updated scores.

Scoring Output

Predicted Score: represents the likelihood a customer will make a purchase in the next 30 days.



Churn

Scoring Criteria

In order to be scored by the Churn model, a customer must have made a purchase in the last 365 days.

If a customer does not fit this criteria, they are dropped out of the scoreable population and will not receive updated scores.

Scoring Output

Predicted Score: represents the likelihood a customer will not make a purchase in the next 180 days.



Lifetime Value (LTV)

Scoring Criteria

In order to be scored by the LTV model, a customer must have done one of the following actions within the last 365 days:

  • Made a purchase
  • Clicked an email
  • Triggered a web event

If a customer does not do this, they are dropped out of the scoreable population and will not receive updated scores.

Scoring Output

Predicted Score: represents the likely total amount a customer will spend in the next 365 days.



Next Purchase

Scoring Criteria

In order to be scored by the Next Purchase model, a customer must have made their first purchase in the last 14 days.

If a customer does not fit this criteria, they are dropped out of the scoreable population and will not receive updated scores.

Scoring Output

Predicted Score: represents the likelihood a customer will make their 2nd purchase in the next 90 days after their first purchase.



Product Recommendation

Scoring Criteria

A customer must have made two distinct purchases within a 365 day window for the transactions to be eligible for this model.

If a customer does not fit this criteria, they are dropped out of the scoreable population and will not receive updated scores.