Case studies articles, page 2
What if your experiment is lying to you?
Before running a Reddit geo test, we checked the smallest lift it could reliably detect. A standard power formula and our simulations disagreed (22% vs 26% in one historical window). We first blamed skewed purchase data. The investigation pointed elsewhere…
Read the article →Want this thinking applied to your account?
Book a free strategy call →