You work on a consumer product where new users see an onboarding message during their first experience. The team believes the message will help users understand the product faster and become more engaged, but it is not clear whether any observed change would reflect real behavior change or just noise.
How would you measure whether the new onboarding message actually changed user behavior? Walk through the experiment you would run, what success metric you would use, and how you would decide whether to launch it.