Tell me about a time you used A/B testing or controlled experiments to make a big product decision. How did you design the experiment and how did you interpret the results?