IMG_4333

A New Copy!

Today, Chapter 1 of  Daniel S. Vacanti’s Actionable Agile Metrics for Predictability. Chapter 1 is titled Flow, Flow Metrics, and Predictability.  Vacanti jumps directly into the deep end by suggesting a way to answer the age-old question, ”when are you going to deliver?”  Buy your copy today and read along!

Flow, Flow Metrics, and Predictability

The first question your customers ask is when will they get what they want.  Customers have an expectation that software-centric team will be able to tell them what they will do and then deliver to that pattern.  Simply put, customers have an expectation that teams are predictable. Historically, software teams and organizations have tried all sorts of techniques to address predictability while failing to address one the critical root causes.  

Most teams are caught in a self-inflicted denial of service attack generated by starting more than they are completing.  Most teams are blind to the problem because traditional project management approaches perpetuate the problem of unpredictability by measuring the wrong things.

A better approach is to measure flow.  The book defines flow as the movement and delivery of customer value through the process. We have seen a similar definition in the re-read of The Goal. Vacanti points out the reason that any knowledge of work framework exists is to deliver value to the customer.  If the goal of any process or framework is to maximize the delivery value the process needs to be optimized for flow.

The size and rate of change in queues indicate flow (or lack of flow). Queues represent work sitting waiting to be acted on by another step in the process. For example, if four developers worked on coding features and then fed their work to a part-time tester, then features waiting to be tested would form a queue.  Predictability, indicated by being able to state when a piece of work can be completed, is directly measured by the size of and consistency of work queues.  In our earlier example, it would be very difficult to predict when an individual piece of work would complete the coding and testing process. Any time the queue is not measured and managed it will be impossible to determine when any specific piece of work will be done.

Measuring queues is the first step in determining whether a team or process is predictable. Measuring queues provides the information to determine when and if work gets stuck waiting. Not measuring queues generates a form of blindness that prevents the action from being taken to get work unstuck.  Recognize that in a very lightweight fashion, the discussion of blockers in daily stand-ups/Scrums is a form of queue measurement.  While recognizing when we need to get work unstuck is important, it would be better if work doesn’t get stuck in the first place.

In Chapter 1, Vacanti defines three metrics that measure queue; therefore, provide data that helps a process or team to become more predictable (assuming the data is leveraged). The three measures are:

  1. Work in progress – the number of items being work on at any point in time (think work started but not completed).
  2. Cycle time – how long work takes from being started to being completed. Cycle time is generally expressed as an average (the higher the standard deviation the less useful the average is to predict when work will be done).
  3. Throughput – how many pieces of work are completed per unit of time.  Velocity is a type of throughput metric (the problem with velocity and story points are tackled later).

Vacanti does a deeper dive into each of the metrics in later chapters.

Metrics are only valuable if decisions are made based on the data and information they generate.  To be useful for decision making, metrics must have predictive power. On page 16, there is a quote from Eric Reis (Lean Startup), “If a metric does not offer predictive power, then capturing that metric is a waste of time.”  Flow metrics replace many of the classic metrics, including classic Agile metrics, like velocity, that are queue blind.  Flow metrics respect person doing the work by not continually front-loading work which shifts us back to pressure to start everything and rarely finish anything

Previous Installments:

Week 1: Introduction and Game Plan

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

Book

Kindle

Get your copy and begin reading (or re-reading)!