Determining decision model readiness using shadow tests and acceptance tests

You want to make a change to your decision model in production — an updated constraint that increases shift time or a new value function that maximizes fleet utilization — but you don’t want to negatively impact business KPIs or operations.

Acceptance testing and shadow testing are useful experimentation tools for determining the readiness of promoting an updated model to production environments. Acceptance testing provides quick insight into go/no-go decisions across specific business KPIs. Shadow testing provides insight into model stability using production conditions and data, but without the production system impact.

The Nextmv platform makes both of these key testing types readily available for algorithm developers and product managers to create and run via a unified experience. Join us as the team behind these two capabilities talks through how to use these two testing types, demonstrates them, and answers your questions live. Please note that we have two available time slots for this techtalk - make your selection on the registration form.

Key topics

  • Overview of algorithm testing framework
  • How to approach acceptance and shadow tests for decision algorithms
  • Live demos of these features and Q&A with the creators of these features

Get started on Nextmv for free and learn more in the documentation.

Have questions? Reach out to us to talk with our technical team.

Presented by
Decision scientist
Product Manager