Question

A retail company deployed a demand-forecasting model on Vertex AI six months ago. The team observes that prediction accuracy has gradually declined over the last two months, but Vertex AI Model Monitoring reports no significant feature drift or prediction drift alerts. The input feature distributions closely match the training baseline. Business stakeholders confirm that customer purchasing behavior has shifted due to a new competitor entering the market. What is the most likely cause of the performance degradation, and what should the team do?

Accepted Answer

This is concept drift; the relationship between inputs and the target has changed even though input distributions are stable, so the team should retrain on recent labeled data.

Answer

This is training-serving skew; the team should verify that the serving preprocessing code matches the training transformation logic.

Answer

This is data drift; the team should lower the drift-detection threshold in Model Monitoring to catch the shifting feature distributions.

Answer

This is a serving infrastructure issue; the team should increase endpoint replicas to restore prediction quality.