How Kanban Worksoriginally written by Amr Noaman Abdel-Hamid
Our summary and key takeaways
If you're wondering what exactly makes Kanban effective, whether it's the process visualization and task tracking, the WIP limits or the ability to measure and improve performance - read this article to find out what Amr Noaman Abdel-Hamid's research found about it. He performed a study of Kanban with regards to queuing theory and Little's Law. In software development Little's Law boils down to an relation, in which Work In Process equals Throughput times Cycle Time. Here, presented are 3 of the most common problems associated with Kanban application in process management.
Problem 1: Increasing throughput
In a situation, when a team's workload rises, but the team size and capacity remain static, there are two solutions stemming from Little's Law equation: reducing the cycle time or rising the WIP limit. With a non-changeable team size, the cycle time cannot be changed, so what seems to be left to work with is the WIP limit. Now, this will work but only up to a point.
As you can imagine, once the WIP limit is increased, the team will start to work harder and most likely to reduce the number of wasteful and ineffective tasks, until reaching a top productivity level. At this point any further WIP limit increase will begin to damage both the productivity and the team itself. After all, there is only so much one can do before going crazy. Therefore, the best solution that we can come up with is yes - increase the WIP limit some, but once you've reached the peak, begin employing more people! If this is not possible, the only solution is decreasing the cycle time, which will mean that items will be delivered more slowly, but your team will stay sane!
Problem 2: Balancing the cycle time
There are cases, where the number of requests to process is unstable from one day to another. This produces a fluctuating cycle time, indicating that at some times the team cannot process the work in time and other times their capacity isn't met. Following Little's Law, an answer to this problem is balancing both WIP limit and the team capacity. Rising the WIP limit will increase cycle time and involving more people will decrease it. Hence - by optimizing these two (WIP limit and team size) will help maintain an optimal cycle time.
Problem 3: Putting a stop to over prioritizing
It's understandable that some items will need to be pushed to the front of the queue, as they may concern vital and urgent bug or unsatisfied customer requests. The danger here is improper balancing of a fast lane on you Kanban board. Should the amount of tasks in it start to grow (or even remain steady), having to attend to these priority tasks may cause a very unproductive dynamic. Having to switch to expedited items means waste in a form of switching tasks time, stopped work in progress on anything that was being done and an overall decrease of throughput. Ideally, an emergency lane is only good for this exactly - emergencies. It's unadvised to keep it on the board in general, as people will always try to prioritize some tasks, meaning the process will get damaged.
When applying Kanban, there are 3 main issues that the team manager should be aware of:
- increasing WIP limits for a team of steady capacity will be beneficial only up to a point
- to balance the cycle time well, keep the WIP limit and the team size flexible
- keep the high priority lane open for emergencies only, as letting it become an everyday occurrence will destroy the process as a whole.