View a markdown version of this page

Amazon AppConfig experimentation - Amazon AppConfig
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Amazon AppConfig experimentation

Application experimentation is the systematic process of testing hypotheses on software applications. It uses A/B testing, feature flags, and analytics to measure user behavior, optimize performance, and drive data-driven product decisions. Experimentation enables teams to validate changes, such as new features or UI changes, before a full rollout, thereby reducing risk and improving engagement.

Built with the expertise of 25+ years of software experimentation in Amazon (including on Amazon.com), Amazon AppConfig experimentation integrates seamlessly with your existing Amazon AppConfig workflows, providing a comprehensive solution for feature management and experimentation in cloud-native applications. Amazon AppConfig experimentation offers the following features:

  • Easy experiment setup and management through the Amazon Web Services Management Console.

  • Powerful and flexible segmentation options for experiment participants.

  • Best-practices for data science analytics between control and treatment groups.

  • Data portability – use your existing data warehouses and analytics tools for experiments or use an Amazon-native service like CloudWatch.

Note

Use Amazon AppConfig experimentation on a pay-as-you-go basis. You are billed for experiment hours run. For more information, see Amazon AppConfig pricing. (Amazon AppConfig is listed as a feature in Systems Manager.)