 A New Copy!

Today we tackle Chapter 5 of  Daniel S. Vacanti’s Actionable Agile Metrics for Predictability. Chapter 5, Flow Metrics and CFDs combines the concepts of flow metrics and cumulative flow diagrams (CFD). Cumulative flow diagrams are powerful data visualization techniques, combined with flow metrics the power becomes more evident. Buy your copy today and read along!

Chapter 5: Flow Metrics and CFDs

Vacanti opens Chapter 5 with a reminder that the reason he was so pedantic in Chapter 4 about how to construct a cumulative flow diagram (CFD) is that mistakes and shortcuts in the rules he previously laid out will make it difficult if not impossible to use the analysis technique described in Chapter 5.

As a reminder, the three flow metrics are:

1. Work in Progress (WIP)
2. Average Cycle Time (in this chapter – approximate average cycle time)
3. Average Throughput

As with the earlier chapters, Vacanti calls out specific properties explicitly. Building on properties in Chapter 4, the first CFD Property in Chapter 5 is:

CFD Property 3:  The vertical distance between any two lines on the CFD is the total amount of work that is in progress between the two workflow steps represented by the chosen lines. In other words, the vertical lift equals work in process.

An example of the property: The work in progress on August 25, 2017, is 21.  10 Units are done, 8 are in QA and 13 in development.  Visually we could ascertain the WIP for any point in time or for any part of the process based on the CFD. WIP Example

CFD Property 4: The horizontal distance between the top and bottom line at any point in time is the approximate average cycle time.

An example of the property: In the example the approximate average cycle time between August 21 and August 25th is 5 days.  The formula is (8/25 – 8/21) + 1 Day. The rationale is that the day the work is started should be counted. In the example we can see an issue, the bands flatten out which quickly after the start of the time period.  When the bands flatten the approximate average cycle time increases, suggesting a problem. Approximate Average Cycle TIme

Vacanti explains that the average cycle time read from the chart is approximate because the items counted as beginning and end at any point might not be the same.  In our example, ten units start and ten complete, but nothing suggests they are the same items.  If we are following the assumptions for Little’s Law (Chapter 4) the approximation will be good.

CFD Property 5: The data displayed on the CFD only represents what has happened not a projection of what will happen.

Vacanti completes the section on approximate average cycle section by pointing out that comparing the approximate average cycle time to the exact actual average cycle provides insights into the process to be discussed in Chapter 9.

The third flow metric, average throughput is defined as the slope of the bottom line between any two points in time on the CFD.  The line represents both the average throughput and departure rate.  Remember slope is rise over run.

In our example, the throughput between August 24th and 25th is 5.

CFD Property 6: The slope of the bottom line of the CFD is throughput as established by departure rate while the slope of the top line reflects the arrival rate of work to the process.

The relative simpleness of the CFD belies the power of the chart.  All three flow metrics can be visualized using the CFD (assuming the data used to build the CFD follows Little’s Law).  By being able to quickly see WIP, approximate average cycle time and average throughput teams can identify potential issues as they are occurring rather than waiting until the end of a time period to notice that you haven’t delivered.

Note: Pages 106 and 107 include all of the six CFD properties summarized in one place.

Previous Installments

Introduction and Game Plan

Week 2: Flow, Flow Metrics, and Predictability

Week 3: The Basics of Flow Metrics

Week 4: An Introduction to Little’s Law

Week 5: Introduction to CFDs

Actionable Agile Metrics for Predictability: An Introduction by Daniel S. Vacanti