- They treat the company like a product – continuously iterating the way in which they work, always looking to improve it in new ways. They've evolved their way of working over a decade of refinement
- They work in six week sprints (although he doesn't like the word 'sprint'), believing that a good version of most everything can be accomplished within six weeks:
'We believe there’s a great six week version of nearly everything. Occasionally some things fall outside of this limit — deep R&D projects, brand new tech we’ve never used before, etc. But we’ve come to discover that nearly everything important can be done in six weeks or less. And done well.'
- Scope is reduced or reconfigured until it fits into six weeks, which sometimes forces them to figure out the true essence of a feature ('Not what can it be, but what does it need to be?'). The six weeks is all implementation and action, with a gap between sprints to pick up on unscheduled stuff, allow for context switching, and planning.
- They split dedicated resource over those six weeks towards both 'big batch' (big features or things that will likely take the full six weeks), and 'small batch' (tweaks, amends, smaller updates that might take from a day to a couple of weeks to complete). They take in one or two big batch projects in a six week cycle, and between four and eight small batch.
- Each big batch project is assigned a separate team, small batch projects are all done by one team. Teams are comprised of two to three people ('complexity begins to increase exponentially beyond that'). They stay together for the full cycle, but come together through a combination of self-organisation and assigment against preferences. Teams often change up after the cycle, but sometimes stick together over several cycles
- Everyone tracks work in the same place, communicates in the same place (Basecamp of-course – big batch projects get their own project, small batch are housed within a single project managed by a single to do list). They don’t (where possible) wait until the end to start QA so there are no bottlenecks
- There is no distinct time set aside to come up with ideas, nor a distinct group of people: 'Ideas are always in motion. There’s always a bubbling ocean of ideas. Every once in a while a bubble floats out of the ocean and lands on the shore. That’s when we begin to take a closer look.'
- When an idea feels formed enough, it’s turned into a pitch ('a fully-formed definition of the problem as well as a proposed solution'). In much the same way as Jeff Bezos believes a powerpoint presentation obscures the true context/detail of a business case and so insists on written pitches, they write up pitches and post them for review as they believe that this enables more effective consideration
Always changing, never standing still, working in blocks of time even for the largest projects, dedicating resource to both large scale and small change request so that neither is ignored, small-multidisciplinary teams, reflection and planning time, robust ways to harvest ideas. As I say, lots to like.