CrazyLabs is a hypercasual and casual publisher whose games have been downloaded more than 4 billion times. They run popular titles, such as Super Stylist - Makeover & Style Fashion Guru, Tie Dye, and Phone Case. Their business model relies on identifying potentially high-profit games early on among many applicants and helping them scale.

At any point in time, CrazyLabs can have up to 15 games that they are testing in order to maximize revenue without decreasing user engagement. Their tests involve running up to 30 configurations of ads plus in-app purchases setups to identify early signals for high potential revenue. They then use these signals to maximize lifetime value for different audience groups. They also test game level design to see what keeps users most engaged. With up to 30 configurations to test per title and multiple geographic groups, it's complicated and time-consuming to activate so many tests and find high-performing configurations. CrazyLabs had tried internal solutions and other third-party tools, but they were slow, unreliable, or not compatible with mobile apps. CrazyLabs needed a way to automatically perform their tests at scale, in a very short amount of time.