This is the last in a series of posts on the four assumptions behind Little’s Law. If you haven’t read those previous posts I encourage you to go back to understand the background. As a reminder, the four assumptions are listed below.
Four Assumptions of Little’s Law
- The average arrival rate is equal to the average departure rate.
- All work entering the system will eventually depart the system.
- The average age of work remains constant, neither increasing nor decreasing.
- Consistent units are used to measure WIP, Cycle Time, and Throughput.
Today we will briefly discuss the last assumption behind Little’s Law and close by discussing what you can do today to make your system more predictable.
Consistent units are used to measure WIP, Cycle Time, and Throughput
This assumption is perhaps obvious, yet it is still worth discussing. We must use consistent units in each of these measures. While using inconsistent units does not describe any obvious lack of predictability in the system it does result in measures that are impossible to analyze using the assumptions of Little’s Law. For example, if we are measuring WIP in number of cards per day we must measure Cycle Time in days and Throughput in number of cards. If we instead measured throughput in number of story points as is common for Scrum teams we will be comparing apples and oranges when considering the effectiveness of the system.
Conclusion
The most important (and difficult) question that most teams are asked is “When will it be done?” While there are statistically sound methods of forecasting that can help you to answer this question, the ability of a team to adhere to these assumptions is critical to those methods working. While all real world systems are bound to stray from these ideals, there are real steps we can take to help us adhere to them more closely. For those reader’s familiar with Kanban systems you may recognize many common policies as some of those steps:
- Strictly controlling WIP prevents us from starting more work than we finish thereby forcing our average arrival rate to equal our average departure rate.
- Delaying commitments to the last responsible moment and thereby starting work as late as possible helps to ensure that work remains valuable until it is finished.
- Focusing on the oldest work first along with actively removing blockers to finishing work helps to ensure that work is not aging unnecessarily.
Each of these policies also supports the others:
- Controlling WIP helps to reduce unnecessary aging of work due to splitting our focus across multiple items.
- Reducing unnecessary aging helps to reduce cycle times thereby further reducing the chances that unfinished work loses its value before being finished.
This is part of the reason that a Kanban process is often referred to as a system. The parts all support each other to make a better whole. What other policies might your team follow to help them more closely adhere to these assumptions of Little’s Law and thereby improve their predictability?
See also:
If you want to learn more about this topic consider signing up for this upcoming public training class Applying Metrics for Predictability on October 16/17!