One of the strongest books I read this year. Well written and to the point, it dives deep into motivations behind the elements of Scrum.
While the book focuses on “pure” Scrum, many of its ideas are applicable beyond Scrum as well.
Scrum is a way to build fast-moving, cross-functional teams. Some of the basic elements are:
- Team performance is the main focus. Measuring individual performance is not a core part of Scrum.
- Gain commitment and speed by time-boxing. Time-boxing is the practice of assigning work to a specific iteration. Every iteration has an equal, fixed duration. In contrast, Kanban achieves commitment and speed by constraining work in progress. More on that my next post 🙂
- Relative estimation is the preferred method of sizing up work. The core idea is that relative comparisons of task complexity and size are less error-prone than estimating effort in hours.
- Teams determine their capacity by looking at historical performance, called velocity.
- Commitment to continuous process improvement is a core part of Scrum. Every iteration should yield ideas for further improvement in the process. These small improvements are called kaizen.
- Scrum measures team happiness by the end of each sprint. Happiness is a leading indicator of future performance. According to Andy Grove, tracking and evaluating leading indicators is one of the most critical activities a manager should do.
- Road maps are loose visions at best. The actual plan changes from sprint to sprint based on learnings from past sprints.
Further reading: Microsoft published a research paper that discusses practical concerns of implementing Scrum in software engineering teams. It’s a good read.
Shuhari 守破離. Breaks down mastery of Japanese martial arts into three stages.
- In the first stage (守), you obey rules without question.
- In the second stage (破), you experiment with breaking the rules and making them your own.
- In the final stage ( 離), you become the physical manifestation of martial arts ideals, rendering the rules irrelevant. This is an amazing way to think about learning any kind of system, including Scrum.
Cross-functional teams. Small autonomous teams, also called feature teams, are the best way to organize work if you want to move fast. Instagram recently moved to this model. There are a few reasons for this kind of team:
- Hand-offs between teams are an opportunity for disaster. Hence you need your feature teams to be cross-functional
- The goal is set outside the team, but only the team decides how to achieve it
- The marginal value of adding the 10th team member and above is negative
Use Planning Poker for effort estimation. Planning poker gets the best of the team’s collective wisdom when estimating. At the same time, it helps with preventing herd mentality and can surface diverging perspectives on implementation.
Process improvements are a part of the process. Kaizens are a fantastic idea for two reasons. First, by making process improvements a core part of the process they’re improving, you can make sure that they get priority. Second, by making it explicit that everyone needs to come up with kaizens, you increase the chance of soliciting good ideas from teammates.
The product owner needs to have enough time for the role. Jeff Sutherland recommends against appointing senior executives to be product owners. The reason is that a product owner is a time-consuming role. Senior executives tend to be time-constrained, so they might not be the best fit.
If you like the article, subscribe to email notifications to not miss future blog posts like this one!