Cumulative flow diagrams

• How can the cause of delays in the value stream be identified?
• Where are user stories being blocked?
• Are non-value adding activities slowing progress?
• How can the value stream's throughput be optimized and made more stable and predictable?

All of these questions, and more, can be answered by making effective use of Cumulative Flow Diagrams (CFDs). this can be inferred with flow diagrams that cumulatively aggregate time spent and the number of user in the various stages of the value stream (columns in the Kanban board ).

A Cumulative Flow Diagram (CFD) is a stacked display of the number of user stories in each stage of the value stream over time. The Y axis is the number of user stories and the X axis is time. It provides an easy to understand visual of the health of the value stream.

CFDs may be used to identify and remove the causes of bottlenecks (delays), and calculating lead time, cycle time, and throughput. All of which may be used to optimize the throughput of the value stream.

Interpreting Cumulative Flow Diagrams

• Each band represents a stage in the value stream and corresponds to the columns in the Kanban.
• The top band is the backlog or the first step in the value stream if the backlog is not included in the CFD.
• The height of each band is determined by the number of user stories (WIP) in that stage over time.
• A bulge indicates a large number of WIPs in a band at a given point in time.
• Bulges indicates the throughput is blocked or constrained by the succeeding band or stage (the band below the bulge) in the value stream.

The causes of bulges are identified and analyzed to determine how best to increase throughput, for example:

• Increase resources working in the blocking band
• Determine and adjust WIP limits for each stage
• Improve the ability to complete stages in the value stream through automation and/or different approaches.
• Use value stream analysis to identify and reduce or eliminate non-value adding activities.
• Use Little's Law to determine throughput and track changes in throughput over time.