While I’m at the XP2010 in Trondheim, I try to update my blog with some of the interesting sessions I attend. This is the write-up from David Anderson’s keynote speech on Building a limited WIP Society in your Organization.

David Anderson explains that extreme programming is a very Lean method since it is about eliminating waste. He introduced three types of wastes from the Toyota production system: Muda, Mura, and Muri. He claims that XP addresses Muri by skilled craftsmanship that lets a story flow without handoffs. XP addresses Muda as planning, coordination and delivery are lightweight and partly automated. He referred to Joshua Kerievsky, Arlo Belshee and Jim Shore, who sought to reduce the waste available with XP by techniques such as Naked Planning, Agile Workcell, elimination of planning, and Limited Red. The motivation for these changes Anderson claims is to eliminate waste in XP. The same holds for the introduction of Kanban.

He continued that extreme programming has not been for everyone. This lead to organizations resisting to adopt Agile at all. The concept of Lean has been around since 1993, and still there are few people actually doing it. Anderson claims that lots of companies use Lean for the adoption of an Agile method, but very few actually can show their implementation of Lean. While Mary Poeppendieck’s books have been popular, you never meet anyone who is using it. Mary’s books mostly served to explain and justify Agile methods in Lean terms. This is fine, but it doesn’t really help in the adoption of Lean. Managing flow is still the biggest impediment to the adoption of Lean. He concludes, that growing Lean adoption is hard.

The underlying question for him is how to go about introducing Lean when organizations have failed with the adoption of other methods such as XP. The answer is to use a limited work-in-progress pull system in order to catalyze the introduction. This leads him to the explanation of Kanban, and whether it is just a tool.

Anderson’s says, that his motivation for adopting Kanban systems was to encourage an evolutionary approach to change. In developing the Kanban Method he introduced a change management approach. This helps him to emerge Lean in organizations. Kanban creates a complex adaptive system in the workplace from which a Lean(er) way of working will emerge.

The seed conditions of the Kanban Method are

• Visualize Workflow
• Limit Work-in-Progress
• Measure & Manage Flow
• Make Process Policies Explicit
• Use Models to Evaluate
• Improvement Opportunities

Anderson explains not to manage work, but the workflow through the system. He challenges process policies and to make them explicit. A business analyst as a role in an organization is an example for an implicit policy, that needs to make explicit. Anderson shows on a comic example how to apply the previously mentioned conditions in a typical stand-up situation.

Anderson continues to explain three types of models on how to look at problems: Bottlenecks, Variability & Waste. Conversation and Leadership are magic ingredients on this. Leaders need to liberally make the impediments obvious to the team. It”s essential that the team respects the limits set to the work-in-progress.

Anderson explains why he didn’t call the limited WIP society the Kanban society. Referring to The Goal from Goldratt he explains the situation from the hike in the book. He bridges the example from the book to the limited WIP society he created. He claims that any WIP limited, pull system will work. He explains that using Scrum and limiting the work-in-progress will work.

The name Kanban is really a misnomer. There is a growing list of emergent properties in organizations. Process should be uniquely tailored to each project/value stream. Typically no enterprise process definition, and no shrink to fit nor stretch to fit in Kanban. The existing process evolves. A WIP limit on the input queue focuses attention on what to start next. This provokes a focus on value. Sketching an opportunity cost of delay function is easier than asking for an absolute value.

A powerful technique is to allocate capacity across classes of service mapped against customer demand. He shows an examples based on multiply colored cards on the Kanban board where he implemented this technique.

On viral spread he shows five examples which spread their knowledge on Kanban systems worldwide. Anderson showed how merged teams share their resources across swim lanes on a Kanban board.

Concluding, he states that limiting work-in-process can help overcome systemic problems. The limits need to be respected by the team. On the outcomes, these improvements will also improve economic and sociological aspects. What from all of this emerges is an organization that lives all the pillars of Lean: Continuous Improvement, Value Stream, etc. If you want to go Lean, you have to limit your work-in-progress.