Question

What are the data requirements for a successful prediction computation in Segment? What causes the error "PREDICTIVE TRAIT FAILED: Selected target event does not have enough relevant users in the target window with events in the training window"?

Product

Twilio Segment

Environment

Segment Console

Answer

To compute a prediction successfully in Segment, the following data requirements must be met:

• The target event must occur at least once per day across the workspace during the feature window.
• At least 5 different event types must be tracked in the feature window.
• Sufficient historical data is required, typically at least 3 times the length of the target window (e.g., a 30-day window requires 90 days of past data).
• The audience must contain more than 1 non-anonymous user.
• Predictions should be made for fewer than 10 million users; if more, define a smaller audience.
• At least 100 users must perform the target event and at least 100 users must not perform the target event.

If these requirements are not met, prediction computation may fail with errors indicating inadequate data.

Additional Information 

For more details, refer to Segment documentation on predictive traits and data requirements.