That was an exciting conference. Not the least because it was the first time for me that I met in person the Lean and Kanban thought leaders that you can follow on the leanagile and kanbandev mail lists (Mary & Paul Poppendieck, David Anderson, Allan Shalloway, Karl Scotland, and many others).
The conference was about exploring the Lean experience in software development with some input from other domains (construction, services, product development). It was a success (in my opinion) as I got a broader view on Lean and Kanban.
I present here extracts of some of the keynote speeches:
May Poppendieck – The Tyranny of the Plan
Mary took us back to a time before there was any computer and showed us how plans were done. Did you know that in 1930 the Empire State building was conceived, designed and built in less than 16 months? Obviously without any help from a computer to create the perfect plans.
The secret behind the Empire State rapid development is to let your constraints (context) frame your design and design as you go. Today, the building would first be designed completely and then a computer would produce a perfect plan to adjust the production to the constraints. It works, only if everyone do exactly as told any nothing “unplanned” occurs (yeah, right…).
Here are some key take-aways:
* “Design the product after the constraints. Do not come up with the design first and try to make it fit in the constraints later”
* “Establish constraints so that you can do your work”
* “Constraints make you creative!”
* “Workflow is orthogonal to schedule” (This one needs more explanations, in a future post)
* “To succeed you need to break dependencies: both related to architecture and related to schedule ” At least wait until the last responsible moment to commit to a dependency
Jeff Patton – Lean Product Development
Jeff re-explained the old agile principles of iterative and incremental development. I was inspired by his view on how you need to wave in product delivery (solution creation) and discovery (understanding the problem) in each iteration. He attacked the agile principle of having a releasable product after each iteration. His view is that “releasability” of a product is not something you get from the burn-down charts. You need to grade the releasability of each feature after each sprint so that the Product Owner can make timely decisions on when to release.
Jeff also presented a very practical way to slice your stories so that big stories can fit into an iteration:
* First, deliver the bare necessity so that the feature “works”
* Then, increase its capability and flexibility (e.g. alternative flows)
* Then, increase safety (e.g. validation)
* Finally, increase usability, performance and sex-appeal of the feature
Karl Scotland – Kanban 101
Karl held a Kanban presentation for beginners where he covered the Kanban fundamentals. His 5 Kanban concepts are an easy way to get into Kanban.
There are 5 key concepts/principles to Kanban:
1. Map the value stream
2. Visualize the value stream
3. Limit WIP
4. Establish a Cadence
5. Reduce the number of kanban tokens
Other interesting concepts:
* “There is kanban (the actual visual card) and Kanban (the system)”
* “Another name for Kanban could be “Flow-based software development””
* “Cadence is different than timeboxing. Timeboxing is like a metronome – mechanic and constant. Cadence is more like a drummer – the rhythm may change.”
The majority of the people attending did have some Kanban experience and Kanban was regarded as the best approach to succeed. Though, all the persons I met had extensive Scrum experience and Scrum seem to be the default container of the majority of the Kanban implementations.
The conference provided the opportunity to meet the Agile Alliance board. We had interesting and passionate spontaneous discussions (though, after some glasses of wine everything becomes rather passionate) around some white boards in the cellars (“dungeons” would be more appropriate) of the Royal Society for the advancement of the Arts. The subjects touched the actual Scrum versus Kanban debate, how to scale, how to coach and several models (strangely enough all very much v-shaped) about how to scale.
What I get out of the conference are several efficiency-thinking tools that I am eager to try in the real world. I am thinking about organizing these thoughts in some way (some future post).
Overall I had a great time, though not enough to enjoy the usually good London weather.