Sign that says walk you bike beyond this point

Progress requires walking and riding!


The 3Ms, Muda (waste), Muri (overburden), and Mura (uneven) are core concepts in the Toyota production system (TPS). When discussed from a lean or process improvement perspective outside of the TPS, Muda gets the most focus because waste feels more tangible. However, Mura, defined as the lack of uniformity or unevenness in flow is even more central to a discussion of change than Muda. Addressing Mura requires a focus on flow. Focusing on flow provides an oversized impact on process improvement because it affects consistency, predictability and even waste and burden. Improving flow is the keystone for improvement.

Toyota’s focus on flow spawned the concepts of just-in-time (JIT) production systems. In manufacturing and operations, JIT approaches require integrated supply chains that minimize inventories, reduce batch sizes and move closer to continuous delivery. In manufacturing, Kanban is a technique that implements both a flow focus and JIT concepts. We already know that knowledge work – such as software development and maintenance or customer-facing operations – are not necessarily driven primarily by interruption and therefore are not inherently resistant to a focus on flow.

Daniel S Vacanti, author of Actionable Agile Metrics for Predictability (interview and re-read) used a simple measure of work item cycle time to focus on flow. Vacanti’s book suggests that by measuring cycle time, teams and organizations and improve and deliver more value to their stakeholders. Cycle time is a measure of flow. Steve Tendon, author of Hyper-Productive Knowledge Work Performance: The TameFlow Approach makes a similar set of arguments based on the Theory of Constraints (Goldratt, The Goal).

Reducing Mura takes two related paths. The first is reducing variability. We define variability as a lack of consistency. Scatterplots are a tool to visualize the definition of variability. Using the scatterplot diagram developed during the re-read of Actionable Agile Metrics (week 11), variability is the range between the lowest and highest cycle time. Reducing variability would flatten all of the observations on the scatter plot (the range would be less).

A menu of techniques to address variation includes reducing interruptions and introduction of expedited work items, reducing batch and work item sizes, addressing workload, solid definitions of ready and done and backlog grooming. Deciding on which approach(s) will be effective requires an understanding of the team or organization’s context. Enter the need to go and see what is happening at the team level (Gemba Walk!).

A second and related avenue for reducing Mura is to increase throughput. That is, focus on increasing the delivery rate of work items. Involve the team in finding ways to deliver value faster. Faster is not permission to deliver lower quality or to game the definition of done. When challenged to reduce cycle time (increase speed) teams will usually focus on reducing overhead and non-value-added steps (you will recognize this as Muda). Address getting work done faster using many techniques such as shifting testing to the left, using test first techniques, reducing batch and work item size, scaling overhead, reducing non-value-added meetings, implementing continuous delivery and backlog grooming. Again, which technique makes sense depends on context.

Reducing the Mura requires a focus on reducing variability. Cycle time is a direct reflection of variability. Improving cycle time is the cornerstone of any process improvement approach. Reducing the variability of cycle time means that teams and organizations have to address the systemic issues identified as other forms of waste.