The basics of A/B testing are simple: randomly assign users to groups and observe their behavior. But things get tricky when it comes to using A/B tests to make major decisions, many subtle mistakes can lead you to draw the wrong conclusions. Andrea will discuss the evolution of Pinterest’s A/B testing platform, and how you can learn from her mistakes to go from simply running experiments to actually deriving insights.
Andrea Burbank works as a data scientist at Pinterest, where she has led A/B testing for the past 2 years. Prior to Pinterest, she worked as a software engineer at Bing and as a natural language scientist on ranking and relevance at Powerset, a semantic search engine acquired by Microsoft in 2008. She has a BS in physics and a BA in linguistics from Stanford University.