No abstract

No abstract

No abstract

No abstract

Data-driven and continuous development and deployment of modern web applications depend critically on registering changes as fast as possible, paving the way for short innovation cycles. A/B testing is a popular tool for comparing the performance of different variants. Despite the widespread use of A/B tests, there is little research on how to assert the validity of such tests. Even small changes in the application's user base, hard-or software stack not related to the variants under test can transform on possibly hidden paths into significant disturbances of the overall evaluation criterion (OEC) of an A/B test and, hence, invalidate such a test. Therefore, the highly dynamic server and client run-time environments of modern web applications make it difficult to assert correctly the validity of an A/B test. We propose the concept of test context to capture data relevant for the chosen OEC. We use pre-test data for dynamic base-lining of the target space of the system under test and to increase the statistical power. During an A/B experiment, the contexts of each variant are compared to the pre-test context to ensure the validity of the test. We have implemented this method using a generic parameterfree statistical test based on the bootstrap method focussing on frontend performance metrics.