Background
Energizing a kanban system has been an area I’ve been thinking of lately ( see my lkbe11 talk). One of the key areas of focus is how to expedite handling of exceptions to flow. What are those exceptions? Blocked items as well as “struggling” items that are not making the progress you expected. Identifying what is struggling is a non-trivial task especially in the face of variability, some of it inherent to knowledge discovery processes and unavoidable. I blogged some of my thoughts about this some time ago, and a collection of possible techniques is also available in the lkbe11 talk mentioned above. Assuming we have some sort of way to classify work as “is currently a flow exception” we want to leverage this.
Exception Handling – Maybe what the doctor ordered?
Last week I was having a discussion with a team lead about the fact she thinks the team is not focused enough on dealing with exceptions, causing the exceptions to be the norm, the velocity/throughout to go down creating an overall feeling of sluggishness. It is also hard to convince this team to use a physical board. They claim they have the data in an issue tracker. So an idea that came up is to maintain an exception-focused physical board, showing just the items that are blocked or struggling. This will the team to start looking at cycle times, and do some kind of TOC buffer management where after the mean cycle time passed, an item goes on the board. 
We think that visualizing these exceptions as well as having the daily meeting focused JUST on them will help the team, and even more important focus them on process and artifacts that can help them. This might create a chain effect of them continuing to fine tune their process and interactions to help them. I will report about the results of this experiment… BTW This reminds me of something I heard at LKCE11, I think it was Jim benson talking about TLC (The Library Company) – where they visualized the longest living items as a big visible chart, and within minutes the longest living items disappeared from the list. It is just a matter of what you look at… 
Conclusion
As teams venture deeper and deeper into flow land, exception handling in realtime might turn out to be an effective way to improve cycle times, flow, and drive improvement at the team level. Teams should consider explicit policies for how they will handle exceptions to flow – will they swarm? slice more thinly? increase frequency of standup meetings? pull a specialist in and pair program with him? Even the discussion of flow exceptions and the need to do something about them would probably be an improvement in and of its own. I’m looking forward to the ability of more and more electronic issue tracking / kanban tools to provide ability to focus on flow exceptions like the ones mentioned above. Several tool vendors already have some capabilities. Leankit Kanban visualizes due dates in danger and has a filtered view that can probably be modified to see exceptions. Silver Stripe provides custom alerts. Another vendor has already shown me an alpha which is in the right direction. PS If you are a vendor with such a capability, feel free to comment and point to what you have… In general, I’m hoping tools will focus more and more on flow/time rather than velocity/throughput. It is a much more interesting and actionable perspective of the work…