In John Stenbeck’s book “PMI-ACP and Certified Scrum Professional Exam Prep and Desk Reference”, he creates an “agile project management process grid” which describes 87 processes used in agile project management. These processes are divided into five process groups (Initiate, Plan, Iterate, Control, and Close), which are analogous to the five process groups in traditional project management, and seven knowledge areas which can be mapped, more or less, onto the ten knowledge areas in traditional project management.
Work-in-Process (WIP) is a concept borrowed, as much of agile methodology is, from Lean thinking to improve the value stream of product development. The basic idea is that a project is a type of WIP because the business cannot realize value from user stories in process, but ONLY from potentially shippable increments of the solution.
In the first of three posts, I explained how WIP is measured. In the second of three posts, I listed some ideas John Stenbeck gave on how to reduce WIP. One tool in measuring the effective of measures to reduce WIP is called Little’s Law, and that is the subject of this post.
Little’s Law was created by John Little of MIT, and it applies to the relationship between
- cycle time (the time required to deliver a potentially shippable increment of the solution)
- WIP (the number of stories that are currently being developed, but which are not yet at a point where they are potentially shippable), and
- throughput (the rate at which stories are being developed in the system)
Let’s take two examples from p. 182 of John Stenbeck’s book.
- 200 stories of WIP, where the team produces 10 stories per day: here Cycle Time = WIP/throughput = 200/10 = 20 days required to deliver a potentially shippable increment of the solution.
- 200 stories of WIP, where the team produces 100 stories per day: here Cycle Time = WIP/throughout = 200/100 = 2 days required to deliver a potentially shippable increment of the solution.
The reason why these two examples were chosen by John Stenbeck is because the WIP is the same. The variable that changes is the throughput, the number of stories the team can complete in a day.
So if you want to reduce the cycle time there are two ways of going about it:
- reduce the WIP, ways of doing which are listed in the last post, or
- INCREASE the throughput, i.e., the number of stories a team can produce in a day.
This can be done obviously by throwing more people onto the team, but there are some processes external to the team which may be influencing how fast the team can come up with solutions, i.e., clarity of the requirements.
The next process in the “Value-Driven Delivery” knowledge edge that is used regularly in iterations is 2.10 Cumulative Flow Diagrams, which is covered in the next post.
Filed under: Uncategorized |
Leave a Reply