This week we finish our re-read of Actionable Agile Metrics Volume II, Advanced Topics in Predictability with two chapters. Chapter 13 is titled Scaling Monte Carlo. In a perfect world, one team could do all of the work needed for a product or feature a manageable piece at a time. Unfortunately, this is the real world. In Chapter 14, titled A Parting Thought, Vacanti states, “We take action because of our relentless pursuit of improvement. That is what professionalism is all about.”
Chapter 13 – Scaling Monte Carlo
In a perfect world, one team could do all of the work needed for a product or feature a manageable piece at a time. In our perfect world life would be simpler and even calmer. The real world is messier. Even though almost all of the mess is self-inflicted, no one has been able to wave a magic management wand and clean things up. Chapter 13, Scaling Monte Carlo discusses how to use Monte Carlo Analysis in some of the typical messes.
There are several aha moments in this chapter but the first is the simplest. It is the explicit statement of four common assumptions that most simple Monte Carlo Analyses (MCA) make.
1. One Team
2. One Backlog
3. Controlled WIP
4. No Dependencies
Assumption #3 makes me chuckle because stated differently, it is an assumption of disciplined work intake, which is a rare occurrence. The assumption that work intake is disciplined and rare is the reason Jeremy Willets and I wrote Mastering Work Intake. Step back and consider the four assumptions. In the purest state, are any of these normally true? If untrue, how do we compensate? A more important question is, how can we make those assumptions true?
A second aha moment was the discussion the what level of granularity you should use for throughput when scaling. I cannot think of a senior manager who is interested in story-level throughput or velocity. Leadership is more interested in delivering features, epics, and products. In the real world, teams complete stories that become features, epics, and products when combined or scaled. I agree with Vicanti’s statement “Our wholehearted recommendation when forecasting higher-level work items is to use the finest-grained unit of work you can–and by implication the finest-grained throughput time unit possible.” The truth of the statement would be hard to assail; however, the unit you report in and the unit you track and forecast may need to be two separate units. Highly intertwined but related. Teams work and think in stories. Multiple teams might work on separate stories to complete a single feature or epic. Stories may have dependencies. Tracking, coordination, and forecasting have to integrate the nuances of granularity. Leader interest lies in the features, epics, releases, and products. Scaling therefore is walking a tightrope between balancing what leaders want to hear and the complexity of making that happen.
Embracing forecasting epics or features requires we consider whether all pieces of work are independent and whether the teams, teams of teams, and programs involved practice disciplined work intake. Basic Monte Carlo Analysis assumes that work is drawn in priority order from a backlog. When work jumps the queue or there are hidden backlogs – in other words when work intake lacks discipline using Monte Carlo Analysis requires more complex adjustments that include combining staggered analyses. The complexity of work requires complexity to forecast delivery.
Vacanti sews up the chapter stating, “The best answer to the problem of scaling is to not scale.” Unfortunately, real life seems to take the exact opposite perspective.
Chapter 14 – A Parting Thought
Vacanti states, “We take action because of our relentless pursuit of improvement. That is what professionalism is all about.” Improvement has a long history with a strong heritage from the lean and quality movements that have become part of the agile canon. However, just saying that improvement is needed doesn’t always translate into action. Translating words into action requires a bit of theory, some techniques, courage, and elbow grease. If we are professionals, then we are the elbow grease.
Next week we will have a few final thoughts. I am torn between Dynamic Reteaming and The Human Side of Agile as the next book. Thoughts?
Buy a copy and get reading – Actionable Agile Metrics Volume II, Advanced Topics in Predictability.
Week 1: Re-read Logistics and Preface – https://bit.ly/4adgxsC
Week 2: Wilt The Stilt and Definition of Variation – https://bit.ly/4aldwGN
Week 3: Variation and Predictability – https://bit.ly/3tAVWhq
Week 4: Process Behavior Charts Part 1 – https://bit.ly/3Huainr
Week 5: Process Behavior Charts Part 2 – https://bit.ly/424O5Wc
Week 6: How Much Data? – https://bit.ly/47GVP24
Week 7: Detecting Signals – https://bit.ly/3SjwfdO
Week 8: XmR Charts and the Four Basic Metrics of Flow – https://bit.ly/48j5AU9
Week 9: Myths and Other Considerations – https://bit.ly/3SvfgVU
Week 10: VoP, VoC, Predictability, Monte Carlo – https://bit.ly/3UGdxzT
Week 11: Different Sampling Methods and 85 Percent or Bust – https://bit.ly/48paPlu
