Fire-drills
They can be effective, when they lead to lasting change
We’ve all been through fire-drills. And I’m not talking only about the Fire Safety Director exercising the alarm system and explaining to us how to evacuate the building. I’m talking about those times when we need to get something extraordinary done, fast, so we drop everything and rally around to get that thing done. In my experience, both Bloomberg and Google thrive on fire-drills. Because those fire-drills demonstrate agility, the capability that the companies have to turn on a dime and focus broad resources on a singular goal. It is a powerful feeling of getting things done.
I believe that fire-drills are needed from time to time. They can be effective. Google+ caused fire-drills, and we got it out. Google Instant caused fire-drills, and we got it out. Sometimes you need a whole team or even the whole company to grab a project and quickly throw it into the past. But I also feel that fire-drills should be kept to a minimum, and I really feel that we should never have to repeat the same fire-drill twice. If we do, then this fire-drill is actually a terrible kind of ‘business as usual’. The way to make sure we never have to repeat a fire-drill is to make changing the way we work a key part of the fire-drill itself.
My team and I were once working on a major turn-around of an expansive product, and I realized we needed a fire-drill to truly change direction. It was disruptive and caused plenty of blow-back from the business because it meant putting a lot of things on hold in order to get sufficient time and energy to focus on fundamentally changing course. We had to burn down the defect list and swarm on key performance issues rather than keep adding features because it was clear to me that product quality and performance were far more critical for allowing us to sell what we had than adding yet another feature to something that was too slow and buggy to use. Even more importantly we needed to make fundamental changes in the way we worked — including implementing automated regression tests, unit tests, smoke tests, and disciplined performance profiling— so that a higher level of product quality and performance became ‘business as usual’ for us going forward.
Please share your comments, questions, and ideas below. And please Follow Scaling Peaks to stay on top of the latest posts. Thanks!
