A/B testing is a data-driven way to optimize your game. You introduce small changes to a sample of users and observe their effects. You ultimately take away the guesswork around design decisions in your game.
The feature is simple really, start by creating two versions of your game; Version A and Version B.
Then, send version A to one group of players and B to the other. The one that performs better is your winner! Once you’re done with your experiment, you can roll out the winning variant to all players. With confidence!