The Rotation: A Process of Improvement with Multiple Experiments

The Rotation is an application that lets men rent unlimited streetwear, for one low monthly price. They offer a massive catalog and allow their subscribers to swap items whenever they want.

‍

Experiment Cycle 1 Set Up

Since they had never run a Lead campaign before, we were starting from scratch. We decided to split the testing campaign into two small experiments to optimize our learnings. For the first experiment, we created & tested a total of 144 variants with $2,500 in ad spend!

‍

Experiment Cycle 1 Results

The first experiment gave us a strong base to build off of, with the top performing ad having an average CPA of $1.33.

With such strong performers, we decided to test wider for round two instead of deeper.

‍

Experiment Cycle 2 Set Up

For our second experiment, we cut the worst performing audience in an effort to learn more about the top performing segments. For Experiment Cycle 2 we created & tested a total of 252 variants with $2,500 in ad spend!

‍

Experiment Cycle 2 Results

The second experiment proved out assumptions from the first as the audience, design, and text type winners remained consistent. The top performing ad in Cycle 2 had a CPA of $1.25! It’s clear that The Rotation should use App imagery, text that falls into the B category of their classification system, and the Hypebeast audience for best performance.

‍

Conclusion

By splitting their experiment into two tests, The Rotation was able to capture deeper & wider learnings than if they had only done one. Plus, they were able to confirm the findings from Cycle 1 by retesting and challenging them in Cycle 2. Lastly, they were able to find positive outliers that they can begin scaling!