A/B testing is a common and fruitful practice, but how do you deem something as successful? What happens when you don’t have enough data points? And worse yet, what happens when you have too many data points?
At Wikipedia, we have a moral obligation and an ethical mission to design for every person, every age, every profession, every background, every location. It’s hard. It’s also rewarding. A lot of data comes into play, however, we follow one motto: Data-informed design, instead of data-driven design. Sometimes the hard numbers lose, but the people win. In this talk, you’ll have an insider’s look into Wikipedia’s A/B testing practices for you to analyze, learn, and copy from.