just because a test yields a “statistically significant” result doesn’t mean it’s significant for the business or for its users. I had been so eager to be “scientific” in my approach that I had completely lost sight of the big picture. I was more focused on something measurable and testable than something that might actually drive the bigger results we needed as a business. Now, I try to start by understanding the size of the opportunity: how many users are actually interacting with this thing, and how meaningful are those interactions? If it’s a really minor thing, then a “data-driven”
...more
