Statistical Methods in Online A/B Testing

Meer over het boek

This comprehensive guide to statistics in online controlled experiments, or A/B tests, makes statistical inference accessible to readers with little prior experience. Each concept is thoroughly explained and illustrated with practical examples from website testing, allowing for immediate application in daily work. It's essential for anyone seeking to understand data-driven business decisions through experimentation, including conversion rate optimizers, product managers, growth experts, data analysts, and marketing managers. The fresh perspective on applying statistics to achieve business goals makes it engaging even for seasoned statisticians. The book introduces scientific methods grounded in business objectives, such as innovating under controlled risk and estimating the effects of proposed actions before implementation. While it includes necessary math and formulas, the presentation is friendly to those with minimal statistical knowledge. Instead of artificial examples like dice rolling, it uses relatable scenarios that mirror real A/B tests. The author critiques the current statistical practices in online A/B testing, highlighting common misinterpretations and the pitfalls of blindly adopting scientific methods without considering the unique aspects of online business. This guide helps readers avoid these errors and align their statistical practices with their business goals.