This month’s book is Measures of Success: React Less, Lead Better, Improve More by Mark Graban.
“This book will show you how to draw more timely, and more valid conclusions based on your metrics, leading to more focused and effective improvement.”
Do you look at single data points each month? Do you sometimes wonder what changed since last month? Do you make decisions based on the change in one data point? Mark Graban explores the commonly wasted efforts and misguided reactions to data and provides a visual answer to steer leaders in the right direction.
In the interest of helping others become better at process improvement, the author shines a spotlight on Walter Shewhart and Dr. Edwards Deming’s Control Charts. Based on the work of Donald J. Wheeler, he uses the name Process Behavior Charts. “Wheeler coined the Process Behavior Chart term, suggesting that the word ‘control,’ while intended to have a benign connotation, nevertheless has ‘baggage’ associated with it.” The book focuses on how these charts provide leaders a better way to make decisions that drive Continuous Improvement.
The Wall of Data
Excel (or the free version Google Sheets) is the defacto repository for the world’s data but viewing walls of data often results in sensory overload. Organizations often pull Key Process Indicators (KPIs) into Scorecards or Dashboards. These can have warning lights or color-coded indicators to indicate potential process problems.
But these types of displays are often misleading. As indicators turn red, does that mean the process is truly worse? Or the opposite, does green mean it got better? It could just be “noise” or the natural fluctuation of a process.
Typical Reactions to Data
Graban points out that organizations often compare single points of data to goals, budgets or to how a metric performed in the previous month. He quotes a typical “VP of sales, ‘Revenue 10% below our target this month’ or ‘Sales are down 5%,’ it’s easy to say, ‘That’s bad.’ But, is such a comparison meaningful?”
Even if metrics are displayed in dashboards with helpful visuals the analysis still falls short. “Leaders might be under pressure to judge performance on a daily (or even hourly) basis. Color coding performance ‘red is bad, green is good,’ can lead to a lot of overreaction, which then wastes the time of managers and their employees. Does any of this help us improve?”
Questions to Ask
Graban points out that “when leaders overreact to every up and down in metrics, it creates pressure on the organization, which might lead to a lot of activity that is more busy than useful. Leadership should be helping reduce waste in the organization, not adding more waste in the name of attempted improvement.” Since the underlying goal is to improve processes, he poses 3 fundamental questions for leaders, with follow-ups, to guide that effort:
- Question 1: Are we achieving our target or goal?
- Are we doing so occasionally?
- Are we doing so consistently?
- Question 2: Are we improving?
- Can we predict future performance?
- Question 3: How do we improve?
- When do we react?
- When do we step back and improve the system?
- How can we prove we’ve improved?
Process Behavior Charts Helps Answer Questions
Instead of comparing something like last month’s hospital readmissions to this month’s readmissions, the answer is to look at data in context. How do this month’s readmissions and last months readmissions fit into the past year of sales? The two charts used throughout the book are the the X Chart, also called the Individuals Chart and the Moving Range or MR Chart. These data displays contribute three invaluable dimensions to analysis:
- Plotting Data Over Time — provides a way to see trends and shifts
- A Center Line — provides a view of the average performance
- Upper and Lower “Natural Process Limits” — show the expected range of performance (aka “Control Limits”)
The trick is to use these charts as a way to view the system as a whole. That way leaders can answer the key questions above. If the process is not consistently hitting the goal, and it’s not improving, that’s the time to go into action.
The combination of viewing data over time, understanding the average and seeing the expected boundaries of performance helps leaders take the right action. If this week’s readmissions are truly off the mark, the X Chart will show that. But if the process is just not capable of hitting the target then either the target was not well thought out or it’s time to launch a proper improvement effort.
The Author’s Key Points
Mark Graban deploys a large array of industry examples to make the case for using Process Behavior Charts instead of what he references as “Bowling Charts” (because they resemble “the grid you’d use to keep score when you’re going bowling”). He showcases the misguidedness of reacting to single data points by bringing whole timelines into view. He makes clear how much time and effort leaders have wasted by reacting to process “noise” and how much more effective it is to guide improvement efforts by understanding true process “signals.” For those unfamiliar with Control Charts it’s a well-crafted argument for turning them into decision-making staples.
As he compares the different methods of data analysis he weaves in ten key points regarding the relationship between data and process improvement.
Summary of Key Points:
- We don’t manage the metric; we manage the system that leads to the results and we lead the people who help us improve the system.
- Two data points are not a trend.
- No data have meaning apart from their context.
- A chart will always tell us more than a list of numbers.
- The job of management is not just to look backward, but also to look forward and predict, if possible, what is likely to occur.
- There is variation in every metric or data set. Process behavior charts filter out noise so we can identify signals.
- Don’t waste time explaining noise in a metric. There is no simple, single “root cause” for noise.
- More timely data is better for improvement. Daily is better than weekly, which is better than monthly, as long as we don’t overreact to every data point.
- If there was an intervention in the system, make it clear in your chart or your discussion of the chart when that change was started or implemented.
- When showing the “before” scenario, show enough data points to illustrate the previous level of variation, not just a single data point.
Support for Using Process Behavior Charts
The author provides an Excel Template for a Process Behavior Chart along with guidance on how to use it. He also includes a chapter on the kind of change management required to move cultures away from reacting to single points of data and toward using Individuals (X) Charts of to drive behavior.
It’s great to see someone showcase these time-honored, incredibly powerful visual displays and make a solid argument for why they are one of the best ways to view process data. He points out, “in an age of ‘big data,’ we are too often drowning in numbers and information.” Process Behavior Charts can pave the way for “sustainable improvement instead of knee-jerk reactions.” That’s a great goal for everyone!
About the Author:
Mark Graban is an internationally-recognized consultant, author, speaker and blogger. He is the Shingo-Prize winning author of Lean Hospitals, the co-author of Healthcare Kaizen, and The Executive Guide to Healthcare Kaizen and he’s written numerous articles on process improvement in Healthcare.
He builds upon an education in engineering and management with practical experience working with executives and frontline employees in multiple industries to synthesize and practice methods including Lean Management, continuous improvement, statistical methods and people-centered leadership approaches.
Mark is a Senior Advisor for healthcare clients with the firm Value Capture. He works as a Senior Advisor to the technology and software company KaiNexus. Mark earned his BA in Industrial Engineering from Northwestern University as well as an MS in Mechanical Engineering and an MBA as a Fellow in the MIT Sloan Leaders for Global Operations Program.
Practical Tools and Concepts Covered:
- Statistical Process Control
- Run Charts
- Control Charts
- Individuals Chart (X Chart)
- Moving Range Chart (MR Chart)
- Process Behavior Charts
- Red Bead Game
- Change Management (ADKAR)
Who Should Read Measures of Success?
- Team Leads
- Lean Practitioners
- White Belts, Yellow Belts, Green Belts, Black Belts, Master Black Belts