LKCE Day 2
It took me one month to complete blogging about Lean-Kanban Central Europe. First day of conference is covered in previous blog post.
Morning keynote: “Science of WIP limits” by Don Reinertsen. I had high expectations for this session, since I had never seen Reinertsen “live” but had read his books and saw videos of his talks. Unfortunately, I had seen this talk before (in InfoQ) so the session itself did not offer much new. However, this was a good reminder of WIP and related issues.
My key takeaways from this session were: (i) Systems have tendency to drift to high-queue states and preventing queues has big impact on system behavior, (ii) even a minimum constraint on WIP improves the system without negative impacts and (iii) decisions about queues (WIP, priority, purge,…) should be based on transparent economical framework .. also make transparent if you don’t know the economics
Session 1: “Lean Meetings” by Jim Benson. He has earlier written book about Personal Kanban and now he is on tour advertising his new book. Jim started by misunderstanding Agile and was very American about bashing anything that was not his idea. Then he presented traditional org.charts and pointed out that silo-organizations were actually very effective during the era where (a) cost of change & cost of product creation were high and (b) cost of information was high. Things have changed, so is need for management, information sharing and meetings.
At the end of presentation I had not much to take away. Some new viewpoints: (i) discussion about status and goal are waste in meetings, these should be known by other means, (ii) Meetings should be shared experiences (e.g. working around same Kanban wall) and (iii) meetings fail the same reason projects fail: pre-planning does not work.
Session 2: “ChangeBan” experience story. I skipped this one after 10 minutes or so, I could not understand the point of presentation. I moved to hear Hermanni’s presentation about “System Conditions”. The video is available here (http://bit.ly/RPaSxx). System Conditions was the topic that had the most profound effect on me. It combined systems thinking with organizational reality and showed how (i) change sustains only if the system is changed and (ii) software development is not only software development.
Session 3: “Portfolio Kanban” by Mike Burrows. I was waiting this session since many people ask about “Large projects” and “managing portfolio with Kanban”. I was a little disappointed; I did not get the point of the presentation. There were three case stories and I did not catch anything regarding “portfolio” Kanban. Key takeaways anyway: (i) spiderweb graph to illustrate “depth” of Kanban implementation, (ii) portfolio Kanban is about alignment between managers / customers / peers / team / you.
Session 4: Lightning talks, short sessions without limits.
The first lightning talk was about “Introducing the change” by Pawel Brodzinski. The trick of the presentation was to handle it through PostIt notes; Pawel pasted tens of postit notes while he was talking. I did not get much from the story itself: main point was (I guess) change is about trust.
Second speech started with annoying and embarrassing Star Wars intro. Talk was “May the Forss be with You” by Hakan Forss. He presented basic Kanban and systems thinking stuff through his trademark Lego pictures. Key takeaways were (i) Human brain needs visual management and (ii) Remember to Collect Data, Go See, Create Habit on Continuous Improvement and Experiment in Small Steps.
Third, and the most interesting, was Benjamin Mitchell “Little less action and a little more conversation”. He challenged Elvis Presley and many others by saying that we actually need to discuss more. He pointed out (very correctly) that Scrum & Kanban provide lot of room to discuss about Doing The Work. Then he listed other topics where a little more conversation would be useful: Purpose, Demand/Need, Work Breakdown, Prioritization, Work Distribution, Team Conversations, Possible Value Deployed.
Finally Gaetano Mazzanti presented a quaint idea of “visualizing” Cumulative Flow Diagram with alternative styles and even sounds. The idea may be more entertaining that useful, but he showed how graphs become more alive by illustrating e.g. age of ticket by different colors and state transitions by sounds.
Session 5: “Double-loop Learning” by Benjamin Mitchell. Again this was a presentation that I had seen before. Still it was interesting and I tried to follow it with fresh eyes. Benjamin presented several models for learning and change, including Double-loop Learning, Ladder of Inference, Mindset-Action-Result. We experimented a little with observation and interpretation. Benjamin’s session was inspiring and I still managed to get several takeaways: (i) Make low-inference interventions, sharing observations rather than advise or judgement, (ii) I see only part of picture and re-framing my approach can change the situation, (iii) make sure you see -> check what others see -> create a full picture from observations.
The conference ended with keynote from David Anderson. He presented very new ideas how to measure Kanban in teams. The keynote started with a very good introduction to Kanban, David has really created a nice 15 min package of Kanban basic idea. Then things turned really odd. David presented question: “Which of several teams should take work from queue? Which team is most suitable?”. He studied the question from financial perspective and created an interesting liquidity model for Kanban: each card movement contributes to “liquidity” of the team and the most liquid teams provide least risky and “correct” throughput.
While the idea was sound and internally coherent, there was a big trouble: I think David was asking the wrong question. It is not the point of “choosing teams” or “finding the most suitable team”. I would ask other questions: why do you have three teams? why do you need to choose? does the organization match the customer demand?