3. What We Need….
• Delivery at speed of change in our markets
• Sufficient predictability
• Alignment of efforts to maximize value to
customers
4. Traditional Approach
Development work is:
• Largely invisible
• Highly variable
Organizations:
• Centralize control
• Fail to take an economic view
• Pursue false economies of scale
• Focus on people staying busy
• Try to remove variability
• Rely too heavily on plans
• Ignore queues
5. Delay breeds waste!
Knowledge is perishable
• Developer & analyst discuss a requirement
- A week old VS. - 3 months old
• Tester discovers a bug
- A day old VS. - A month old
Quality suffers
Motivation goes down
Business opportunities get missed
6. Agile helps, but…
• Dev team focused (suboptimal)
• Coordination and scaling challenges
• Underlying principles at odds with traditional
management assumptions
• Won’t solve all your problems
• Can put stress on rest of organization
• Focus on “velocity” can foster overload
24. Limit Your WIP
Kanban at Scale – A Siemens Success Story , InfoQ, Feb 28, 2014
http://www.infoq.com/articles/kanban-siemens-health-services#anch107610
25. Finding Leverage
• Stop using oversized batches
• Make work visible
• Focus on managing queues
– Capacity utilization is hard to control
– Cycle time is a trailing indicator
• Attack queues as they arise
• Queues have a quantitative cost
• Understand the tradeoffs
26. Constrain WIP at All Levels
• Epics, MMFs
– Eliminate excessive early elaboration
– Drop items as you add new ones
• Put WIP limits on all backlogs
• Manage WIP of shared resources, experts
27. Take an Economic View
• Understand full value chain
• Quantify cost of delay
• Don’t consider $ already spent
• Watch the work product, not the worker
28. Take an Economic View
• Influence small decisions
• Quantify life-cycle profit impact of:
– Product cost
– Product value
– Development expense
– Cycle time
– Risk
29. Manage Variability
• Standardize where it makes sense
– Automated testing
– Continuous integration
– Continuous deployment
• Make batches roughly the same size
• Include technical risk in prioritizing stories
31. Use Cadence To…
• Limit the accumulation of variance
• Provide sufficient capacity margin to enable
cadence
• Make waiting times predictable
• Enable small batch sizes
• Reduce communication costs (i.e. meetings)
34. Summary
• Understand your economics
• Make your queues visible and control them
• Create a process to exploit variability
• Reduce your batch size
• Control cycle time by controlling WIP
– Attend to delays
• Sequence work based on economics
• Accelerate feedback with smaller batches
• Push decision making down (where advisable)
35. Additional Info
• Don Reinertsen’s website (blogs & videos):
http://reinertsenassociates.com
• The 175 Principles of Flow: http://lpd2.com/the-principles-of-flow/
• First chapter of the book:
http://www.celeritaspublishing.com/PDFS/ReinertsenFLOWChap1.p
df
• Al Shalloway: Not Doing SAFe? No Problem. Not Doing These? Big
Problem http://www.netobjectives.com/blogs/not-doing-safe-no-
problem-not-doing-these-big-problem
• Kanban at Scale – A Siemens Success Story
http://www.infoq.com/articles/kanban-siemens-health-
services#anch107610