Measuring for Improvement


Note: Welcome LEI visitors… you can check out my main blog page at, as I hope you'll come back. Here are some other recent posts:


By Mark Graban:

When visiting the “gemba” at a few hospitals in England two weeks back (I saw five hospitals in three days), I heard a great expression at one London hospital:

“We're trying to shift from collecting data for judgment to
data for improvement.”


Even though I subscribe to Lean concepts like “the right process leads to the right results,” my focus on process doesn't mean you ignore results. Process measures are important — having the right measures, in a timely way is critical to any lean implementation.

When I've helped hospital in their initial lean implementations, we go from monthly measures that are often 4 to 6 weeks old to daily measures (or even, for example, laboratory turnaround times that are measured each shift). The right metrics, a small number, meaningful to staff, things they can have an influence on through their process (as opposed to complicated financial ratios) – that's what we put in place.

Measurement is like any tool – it can be used in good ways or bad ways.

Many organizations would focus on data for judgment sake. Which individual is performing well or badly? Maybe the variation is due to the system. I saw one large laboratory measure each phlebotomist's individual productivity every day.

That's OK, if those with low productivity are coached and mentored. If you have “data for improvement,” you treat the people differently than if you're using the “data for judgment.” Having measures shouldn't be used, in the old G.E. style, to fire the lowest 10% in phlebotomist productivity.

A lean leader would ask why their productivity is lower. Are they assigned to a part of the hospital where draws are more difficult? This might require a “gemba” view in addition to just measures. Was the employee not trained well or not trained the right way? There are a dozen questions you should ask before just firing or punishing somebody.

If you're using data for judgment, and people are in an environment of fear, all sorts of bad things can happen — people will fudge the numbers or fudge the system. That's why we need data for improvement… the mindset and the philosophy that goes with the data and the measures, treating people the right way so they can focus on improvement instead of focusing on gaming the system.

What lessons have you learned along these lines? I feel like, sometimes, I could write a whole book on this subject. But, instead maybe read the works of Dr. Deming or the outstanding book by Dr. Donald Wheeler, Understanding Variation: The Key to Managing Chaos, probably the most influential book in my professional life.

Subscribe via RSS | Lean Blog Main Page | Podcast | Twitter @markgraban

Please check out my main blog page at

The RSS feed content you are reading is copyrighted by the author, Mark Graban.

, , , on the author's copyright.

What do you think? Please scroll down (or click) to post a comment. Or please share the post with your thoughts on LinkedIn – and follow me or connect with me there.

Did you like this post? Make sure you don't miss a post or podcast — Subscribe to get notified about posts via email daily or weekly.

Check out my latest book, The Mistakes That Make Us: Cultivating a Culture of Learning and Innovation:

Get New Posts Sent To You

Select list(s):
Previous articleWhen is it time for lean "lite"?
Next articleOne More Request for the Lean Blog Survey
Mark Graban
Mark Graban is an internationally-recognized consultant, author, and professional speaker, and podcaster with experience in healthcare, manufacturing, and startups. Mark's new book is The Mistakes That Make Us: Cultivating a Culture of Learning and Innovation. He is also the author of Measures of Success: React Less, Lead Better, Improve More, the Shingo Award-winning books Lean Hospitals and Healthcare Kaizen, and the anthology Practicing Lean. Mark is also a Senior Advisor to the technology company KaiNexus.


  1. … the old G.E. style to fire the bottom 10%…" I used to work for an organization that adopted that slaveship, oops I mean performance management, method. Probably one of the best methods to drive IN fear and inhibit teamwork I've ever seen. No way would one unit manager even consider compromising his productivity metrics even if it meant helping his internal customer, and when the quarterly earnings per share targets were at stake, everyone knew where the priorities were. Ruthless, callous, toxic. And, no, Jack Welch is NOT my relative.

  2. Fred Taylor went over this ground in 1892 after Henry Townw at Midvale. If you can't measure it you can't manage it.

  3. Jim – if you come back, can you elaborate on your comment?

    Rearden – I respectfully disagree with you, as I believe what Dr. Deming taught, that some important things are unmeasurable, but still have to be managed.

  4. Sorry, I should have explained my comment. Obviously I was trying to be funny, and play off the original quote about statistics. (“Lies, damned lies, and statistics.”) We all know how statistics can be interpreted and or spun to explain any position. Actually, I was agreeing with your comments about how we should use measurements and data. We should use the data we collect to ask more questions, dig deeper. Ask lots of questions. Go to Gemba.

    I have on occasion collected data and then found out the data does not reflect reality. Because Lean is a team sport we have to be careful with the lean measurements we gather. Especially if we gather this data before we involve people in the problem solving process.

  5. I noticed that the author mentioned Dr. Wheeler's book on variation as his most influential book. I enjoyed that book immensely, but am not sure how to reconcile that thinking, with daily metric review in a Lean environment. My experience has shown that we end up reacting to every data point that does not meet goal. In other words, is the daily metric reviewed in a control chart? Or, is it a metric measured against a "goal"? If measured against the goal, this is not at all in the spirit of Dr. Wheeler's book.

  6. Chris – you are absolutely right about not overreacting to every up and down in an hourly or daily measure.

    I've always taught my hospital clients to view data in a Wheeler/Deming/SPC type view.

    I discuss this in my book "Lean Hospitals" as well.

    Control charts are a very important method and mindset. It fits with Lean perfectly. When you try looking for a special cause to every up and down in the process, you waste a lot of management time looking for a special cause that likely doesn't exist.

    Measures, like waiting time or daily lab turnaround times WILL vary somewhat, the natural variation of the process.

    The key learning from Dr. Wheeler is using SPC to separate signal from noise, or special cause from common cause, in our metrics and management data.

    I would be guilty of lean malpractice if we put measures in and the measures were used to blame and shame people.

    Thanks for bringing up the point.

  7. Mark – thanks for your reply. Great to hear your thoughts on the use of "statistical thinking" presented in the Wheeler/Deming/SPC books. In the Lean environment that I am in, we measure the metric against a GOAL. Doing so, does not promote the SPC mentality. In fact, it has promoted the overreacting. Do you suggest that the daily metrics are tracked against a GOAL?

  8. Chris – you make a great point. I think there's a difference between looking at the customer's requirement and just setting a goal.

    An SPC approach looks at the historical mean/average that the process gives you. Let's say the average hospital lab turnaround time for a certain test is 90 minutes. Some days, you'll be above average and some days you'll be below. SPC and Wheeler teach us to not overreact to every up and down.

    If our control limits are 75 and 105 minutes, that's where we want to start looking for common cause, when a daily is lower than 75 or higher than 105.

    If the customer goal is 60 minutes turnaround, we have to find process improvements that would allow us to reduce the average/mean of the process, rather than pressuring people to do better. The current process isn't capable of hitting the customer requirement.

    Some organizations plot real data with just the customer goal. This can be demoralizing if the process isn't capable. Better to chart the mean and statistical control limits — and educate staff and managers about what that means.


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.