Do you know the best way to do A/B testing?

Last updated by Piers Sinclair [SSW] about 2 years ago.See history

A/B Testing is the process of testing different versions of an application on different users to gather empirical evidence to learn which version is better.

Using A/B Testing enables you to get features tested and when used effectively means that a bug will never be deployed to 100% of users. Generally, new features should be tested on 20% of users and rolled out to others once they are reliable.

Video: What is A/B Testing? | Data Science in Minutes

There are several ways this can be done...

Feature flags are a modern way to toggle features for users. They are essentially a little bit of code that can be turned on and off at will. That means you can choose, when features are deployed and who gets them.

Feature flags are often implemented by developers writing their own code. However, there are better solutions today:

Azure Deployment Slots

Azure Deployment Slots are another way of doing A/B testing, you essentially deploy 2 versions of your app and then direct traffic to different versions.

Azure FrontDoor

Azure FrontDoor is an offering that lets developers direct traffic to different versions of an app.

We open source. Powered by GitHub