An Elegant Puzzle: Systems of Engineering Management

by Will Larson

Deep in the data. You should be deep enough in your data that you can use it to answer unexpected questions. This means spending time exploring the data, and the most common way to do that is to run a thorough goal-setting exercise.