| Awesome, I'm happy you enjoyed it! In some sense, it might actually be easier to tell at a startup because what matters at a startup is growth. That can be growth of revenue (if you already have a product to sell) or growth of users (if you need traffic first to sell, e.g. Quora). The highest-leverage work would then be the things that most directly lead to that growth. Sometimes, this will require talking with product managers or salespeople or users to understand what the biggest accelerators or roadblocks would be. So, for example, at Quip, I looked at data on how the product spread within teams and organizations, developed engagement metrics around it, and then built features to move those metrics. Working on some projects that end up failing is perfectly normal. What's important is to front-load as much of the risk as possible and to be explicit about the hypotheses required to make a project successful so that you can validate them early and, if necessary, change course. That's where some of the projects we worked on at Quora failed (e.g. topic groups, an infrastructure rewrite) -- we let ourselves be overly confident about what we knew and didn't invest the time to sanity check our hypotheses. For my book, validation played a huge role. Even before I started writing my book, I had written engineering-related answers on Quora for over two years and started to get a sense of what resonated with readers. That helped me build confidence that there would be demand for something in this space. Continued feedback and reviews during the writing process built on that confidence more. |