The Power of Online Experiments – What Big Tech Can Teach Us About Testing Ideas
Article Title: Statistical Challenges in Online Controlled Experiments: A Review of A/B Testing Methodology Authors & Year: Larsen, N., Stallrich, J., Sengupta, S., Deng, A., Kohavi, R., Stevens, N.T. (2024) Journal: The American Statistician [DOI: 10.1080/00031305.2023.2257237]Review Prepared by David Han What’s an Online Experiment & Why Should We Care? Have you ever wondered how Netflix decides what thumbnails to parade on your screen, or how Amazon chooses what items to nudge into your recommendations? Behind the scenes, many of these choices come from something known as A/B testing, a type of experiment that companies rely on to make smart, data-driven decisions. Picture this: After creating two versions of a webpage, you deploy each site to separate groups of users online. You then patiently wait to see which one performs better – say, attracting more clicks or garnering more purchases. This is an A/B test in action. In the tech world,…