You work on a mobile product team running an A/B test for a new in-app feature. Early results look promising after a few days, and stakeholders are asking whether you can call the experiment early and make a decision. You are concerned about peeking and making a premature call based on unstable results.
How do you avoid peeking or making premature decisions during an experiment? Explain how you would set up the test and decision process so that you do not stop early for noise while still protecting the user experience.