Mark's note — today's guest post is by my friend Karen Martin, my guest for Podcast #151 where we talked about her latest book The Outstanding Organization: Generate Business Results by Eliminating Chaos and Building the Foundation for Everyday Excellence
Think you know where PDCA (plan-do-check-act) came from? Think again. As I was writing my book, The Outstanding Organization, I dug into this powerful approach for problem solving and was surprised by what I learned.*
The two most influential voices for fact-based management in business have been Walter Shewhart, who is often referred to as the “father of statistical quality control,” and W. Edwards Deming, the “father of the Total Quality Management movement.” In the late 1930s, Shewhart converted the traditional view of mass production as a linear, three-part model involving specification, production, and inspection into a continuous-feedback quality control loop that relied on the scientific method of forming a hypothesis, carrying out an experiment, and proving or disproving the hypothesis. As Shewhart's editor, Deming helped Shewhart simplify his concepts for publication. Deming continued to modify Shewhart's cycle and in 1950 presented the concept at a lecture to the Japanese Union of Scientists and Engineers (JUSE). The cycle became known as the Deming Wheel: design, production (make), sales (sell), research (redesign through market research). Testing was implicit in all four steps of the cycle.
According to Masaaki Imai, who is often referred to as the “father of kaizen,” an unnamed group of Japanese executives recast the Deming wheel into the PDCA cycle—plan, do, check, act—a four-step cycle for improvement.
Kaoru Ishikawa contributed several key subcomponents to the PDCA cycle, notably, that the plan phase includes setting clear goals and targets, the do phase includes significant training and education during the implementation of any improvement, and the check phase includes continuous revision of standards in response to the voice of the customer and changing requirements of the next step in the process. Through Ishikawa's influence, PDCA began being viewed as a significant management tool and the standard for making continuous improvements to processes and systems.
In the 1980s, Deming expressed his view that owing to translation difficulties from Japanese to English, the PDCA cycle had been corrupted. Deming recommended replacing PDCA with PDSA—plan, do, study, adjust—which he felt was linguistically closer to Shewhart's original intent. Deming continued to refer to the cycle as PDSA through the 1990s and dubbed it the Shewhart Cycle for Learning and Improvement.
I'm with Deming. Over the more than 20 years that I've been involved in quality improvement, operations design, and problem solving, I've found that the words check and act in PDCA mask the intent of those steps, which are to study the results of one's experiment and then make adjustments based on the results from the countermeasure(s) put in place to test the hypothesis. People often mistake check for making sure that everyone is following the new process rather than checking the results of the experiment and adjusting the countermeasures accordingly. In addition, people often act without adequate reflection and adjustment.
I still sometimes slip and use PDCA—it's been habituated in my brain and vocabulary for over 20 years. But I'm committed to helping people become proficient problem solvers as quickly as possible, and I've found using “study” and “adjust” helps quite a bit on this front. So PDSA it is.
*While there are many accounts about who developed what, in what sequence, and when, the most complete and well-documented account I've found is a white paper, Evolution of the PDCA Cycle, by Ronald Moen and Clifford Norman. In it, the authors present a view about Dr. Deming's role in the development of the PDCA cycle that varies from popular folklore.
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:
To PDCA or PDSA, that is the question.
For my own organization it is PDSA. Honestly it came down to a coin flip and we chose PDSA from some early examples from the American Heart Institute. Nice background information and now I will have a more concrete response to why we don’t use PDCA.
Good choice! Deming would be proud. :-)
Helpful clarification. I still use term PDCA, however, because most people regard it as a catchall term meaning continuous improvement cycle. I believe the PDCA/PDSA debate just confuses people.
Example: I coach people on crafting career A3’s. Once the root-cause of what’s blocking them from progress is identified, the “Do” becomes obvious. “C” check is becomes critical when they complete their daily inventory to make sure they are not repeating the same wasteful patterns.
A mindset blocking most careers is the impulse to overanalyze or over study before actually “shipping”. “A” is shipping. Too many folks get locked in “Study” and can’t “ship”. As Seth Godin says, “when you ship things change. Your project interacts with the world and the market changes. You change. Your relationship with your team changes.”
Nothing happens when you study, and study, and study….
Totally agree, Jim. Yet I see an equal number of organizations who “do” and “ship” before they have a clue what the real problem is and what actually caused it. So caution should be used in all four steps.
As for the PDCA/PDSA debate, you can see from my other comment that I don’t necessarily believe changing terminology when one terms already been socialized across the organization is helpful.
I don’t think the term “Study” inherently means to study something forever (ala “analysis paralysis”). I think it means to be thoughtful and not just do a cursory “Check” (as in “check you got the results you expected, which might presuppose success).
A well put together summary of the history and meaning of the Improvement Cycle Karen. I have the conundrum of a poorly understood but embedded use of PDCA terminology in a large organisation. The easier to understand PDSA that puts scientific thinking into context will be a struggle to introduce. I’m not even sure if it is worth trying to change the terminology although the improvement process still requires the correct coaching.
I’d love to hear of any examples where people have unpicked the stitching of PDCA and replaced it with PDSA in a large organisation and how they went about it?
Hello Kopstar – Thank you for your kind words. I’m glad you enjoyed the post. Good question but, unfortunately, I don’t have much experience with changing terminology, especially in a large organization.
I’m not sure it’s worth the effort in “un-learning.” That said, if they’re not embracing PDCA because the words are less clear, perhaps you introduce it in one area and see if it helps. If so, then roll it out across the organization.
If you want to stick with PDCA and want a version of the 12 steps of PDSA that’s PDCA-ified, I’d be happy to send it.
Then main point is to have a consistent approach for improvement and problem solving that’s consistently followed and drives ongoing improvement. You can call it “George” if it accomplishes the mission at hand. :-)
I hope this helps.
Yeah, I’m not sure it’s worth changing the term that’s used if it is well engrained. I’d spent more time making sure the approach is well understood and is consistently and tenaciously applied.
Good post Karen. The reason I prefer -and always use- the word ‘study’ is because I often see the ‘plan’ being downgraded to only ‘who will do what by when’ and ‘check’ subsequently being nothing more than ‘checking that an action has been done’.
‘Plan’ in the meaning of a substantiated hypothesis of how work wil work better, and ‘study’ in the meaning of studying the effects and actual workings when we try to adhere to the hypothesis is often not done.
As you state in the title, I also definitely think words matter. And I personally feel it is worth the efforts to introduce the word to better convey what is really implied.
I also wrote two related posts about this topic that you may want to read here: http://www.dumontis.com/en/blog/broken-pdsa-cycle and here: http://www.dumontis.com/en/blog/pdsa-plan-within-plan.
Thanks, Rob. Agree completely. In my experience, “Check” is the most problematic of the four, but “Act” is right behind it. Thank you for the blog posts. Will take a read. Have an OUTSTANDING day! :-)
Yes, “Plan” is not meant to be a superficial step. As I practice it (and was taught), Plan really incorporates understanding the current state and understanding the problem well. Brian Joiner (in his excellent book “Fourth Generation Management”) wrote that we should really start with Check, as in checking how the current state works. That’s a fine point, but I think that’s really part of an effective “Plan” stage.
Note, there is another “guru” running around claiming he invented the idea of “start with Check” but it’s clearly Joiner who spread that idea (giving credit where credit is due).
Coincidentally, I’ve been wrestling with “Check-Act” vs. “Study-Adjust” for the past several months as I’ve been working with teams and this article has nudged me even more toward Study-Adjust. Yes, it may seem like splitting hairs at times, but “Study” implies more consideration and seems less casual than “Check” and “Adjust” connotes a more temporary situation, which encourages continuous PDCA.
Nice article and discussion.
Oops. Better make that PDSA from now on. Old habits die hard. ;-)
LOL. I suffer from the same occasional slip as well, Mark! Old habits do indeed die hard! Yes, adjust, does do more for creating a CI mindset than Act. Sounds like good timing with the post! All the best!
Karen,
Great post. I wrote about this briefly in my on-line journal (not really a blog). I spent a bit of time with Dr. Deming from 1986-1993. I call it a correspondence relationship (before the era of e-mails and I know Dr. Deming would never send an e-mail if they were in existence back then). I also had some face time with him and I witnessed his reaction to the term PDCA versus PDSA. You would not want to get “the look”. He was always very thoughtful about the words he used (written and spoken). I witnessed some exchanges between Dr. Deming and others (Joiner, Scholtes and others) who tended to use PDCA (I think it was they way they learned it – what Dr. Deming would call an example of “rule 4 of the funnel”). If you want to be true to the principles behind the cycle, the correct word is “study”. A deeper understanding of his cycle for improvement and cycle for learning (he saw it as both) can be found in the book “Mind And the World Order” by C.I. Lewis. It’s a tough read, but an important one.
Mike Stoecklein
Hi Mike. Wow. I would have loved to had face time with Deming. I LOL’d when you mentioned “the look.” I can only imagine! Thank you so much for sharing your perspective and the book tip. I’ll add it to my “must read” list (which is very, very long!).
Hello Karen, thanks for the nice blog. In the article you referenced by Ron and myself, there is a reference (see #15) to a letter from Deming to Ron Moen (who did 75 Four Day Seminars with Deming) where Deming cautioned Ron, “Be sure to call it PDSA, not the corruption PDCA.” This is strong language for Deming. As one of the posts noted, Shewhart and Deming were influenced by C.I. Lewis and the pragmatists as they developed their thinking around the Shewhart cycle. It was not until 1993 that the details of what we know as PDSA came to be with the iterative process of deductive and inductive learning built into the cycle. JUSE has been very clear from the start that PDCA was aimed at implementation of a standard and as Dr. Kano reminded both Ron and myself, this is very different from PDSA. PDSA cycles can be used to collect data, develop changes, test changes and implement changes. These cycles are typically very quick with an emphasis on good testing principles to speed the process of learning. Best Cliff Norman, API
Cliff – thanks for the comment and sharing that story about Dr. Deming.
Hi Cliff – Wonderful to “meet” you on Mark’s blog. I was thrilled when Kevin Cahill at The Deming Institute referred me to your article when I was doing research for my book. It was highly validating to learn that Deming felt that words mattered. A lot. I’d be curious to see what you think about my take on the breakdown of PDSA, which can be found on pp. 121-123 of my book, The Outstanding Organization. Thank you for sharing your insights.
We are a DMAIC-structured shop — always been intrigued by PDSA — how would you “port” the concept to DMAIC structure, if at all?
Thanks, James. I’m curious to hear Karen’s thoughts, but here’s my take. I am Green-Belt trained, but I’ve never really practiced Six Sigma.
I see DMA – Define, Measure, Analyze mapping to “Plan”. Either DMAIC or PDSA should start by understanding the current state, customer value, goals, collecting relevant data, understanding root cause, and brainstorming possible improvements.
I think “Improve” maps to “Do” & “Study.” It should be a matter of using iterative tests of change.
I guess “Control” maps to “Adjust.” One thing that has always bothered me is that I wish “Control” was “Continuously Improve.”
How do you see it?
Hi James – I guess I’ll start by asking if DMAIC is working for you? Is it driving the right behaviors? Are people using it? (interesting that we often say “practice PDSA” whereas many say “use DMAIC.”) If it’s working, there may be no reason to change. Plus, you risk people viewing it as flavor-of-the-month thinking.
If it’s not working, dig deeply and find the root cause. If you find that moving to PDSA is likely an effective countermeasure, I’d probably do it. But be very, very clear on why. You need to have sound reasoning.
Good luck! Please come back and drop a comment to let us know what you decide to do. (You might be interested in my response to Kopstar above.
Ahh… I see I mis-interpreted what you meant by “port,” James. Sorry ’bout that.
Actually I did a mapping exercise with PDSA, TBP, 8D, DMAIC, etc. when I was writing the book. I’ve never publicized it, but perhaps I should, given interest in the subject. Mark, let me know if you want me to give it to you for a follow-up post or if I should blog about it myself.
If you consider the P to include the hypothesis of how work could work better, the P goes all the way from D up to a part of I where you specify the countermeasures. The 2nd half of I is about actually implementing these C/Ms so maps to the D in PDSA. The C is the S and A if done correctly.
My personal experience is that the S (study) and A (act: accept / adjust / abandon), however, are far less open than the traditional C phase in DMAIC. In practice, that often more looks like force-fitting and “making it work whatever” kind of exercise.
Best, Rob
I use both methodologies, PDSA and DMAIC. At a coaching and involvement level I find PDSA and A3 Thinking is a far more inclusive way of involving everyone in change. DMAIC is more exclusive and detailed so therefore difficult to involve everyone. DMAIC tends to be the domain of experts.
You raise a good point. I just got back from a client who clearly needs to use DOE (design of experiments) to solve a pesky multi-variable quality problem. One of the things my meeting highlighted was that we don’t even teach DOE in most Lean programs. I often view Six Sigma as the “big guns” that I need to bring to problems that are big-gun problems, but it does beg the question about why Lean doesn’t include DOE in its toolbox.
Problem solving is problem solving and we need tools that help get to root cause, no matter how complex the problem. Ric on my team’s the one with extensive DOE experience so I’m bringing him into the engagement for this stage of their learning and transformation.
I see Lean thinking and Lean tools as something that can be taught to everyone and applied by everyone.
Six Sigma is a more complicated set of statistical methods that will generally be used by experts. DOE pre-dates Six Sigma, of course, and was something I learned in a college stats course… but I don’t use it at all in my career. I’d probably bring in a “big gun” with experience with that approach, when needed.
I think Lean doesn’t “include DOE” for the same reason the Lean books generally don’t include pareto charts or other basic Q.I. tools… I think the assumption would be that people are smart enough to incorporate other pre-existing tools and methods as needed and that it doesn’t have to be spelled out as some sort of fully formed programmatic approach. The Lean literature generally wouldn’t say to NOT use those approaches.
Agree. But it’s a rare Lean book that mentions root cause analysis as a key part of the P in PDSA/PDCA. That’s a serious flaw and is one reason why is why I keep seeing visual “problem boards” with no RCA. Getting better at defining problems and then leaping to solutions beats an ill-defined problem and then leaping to solutions. But leaping, of course, begets more problems.
I did an experiment with a client recently where I intentionally didn’t teach them any of the simple RCA tools (5 Why’s, fishbone, Parento, etc.). All I did was “require” them to search for the root cause before choosing appropriate countermeasures. Problem solving effectiveness went through the roof. Something to think about.
Hello Kopstar, I found your post interesting. As you probably know, DMAIC is suited for improving an “existing process.” If you need to develop a new process, improve a product or develop a new product then the array of tools list in the DMAIC gates will not apply. Then one of the several DFSS approaches needs to be employed to ensure you have “the correct tools” in the prescribed order. PDSA can be directed to design or redesign processes, products, services or the system as a whole. As Richard Feynman once noted, “Science begins and ends in questions.” So does the science of improvement. PDSA requires the rigor of posing a question that can guide the project. We find this is difficult for many people. Questions are typically phrased in terms of judgment; Will the doctors find the whiteboard useful? A better question; What will be the issues for doctors in terms of using the whiteboard? The first question will elicit predictions of yes/no. The second question will surface the theories in use by the team using PDSA. Dr. Dennett has observed: “…anybody who has ever tackled a truly tough problem knows that one of the most difficult tasks is finding the right questions to ask and the right order to ask them in. You have to figure out not only what you don’t know, but what you need to know and don’t need to know, and what you need to know in order to figure out what you need to know, and so forth. The form our questions take opens up some avenues and closes off others, and we don’t want to waste time and energy barking up the wrong trees.” I believe that DMAIC is attractive to many practitioners because it lays out the tools to be used in some order. Each DMAIC model is somewhat different. We have as many examples as we do organizations or consultants. Each has had a different experience in the use of tools. Using PDSA demands that questions lead to the data necessary to answer the questions and then “only the tools” that are required to collect, display and analyze the data or change. One of my colleagues has challenged me to find one project that has actually followed the prescribed DMAIC model in the real world. I am still looking. Best, Cliff
Thanks, Cliff. If you find one, please let us know :-)
[…] 1 (PDF) has a good illustration of the PDCA cycle (although I, like Karen Martin, prefer to use the PDSA acronym). But, it’s the same thought process, as this diagram shows very […]
Hello Cliff,
Interesting thoughts. I guess the point I was trying to make was that if I’m looking to change/influence the culture of an organisation then I need a simple system or method that can involve everyone. PDSA/PDCA is simple to coach and empower people to go and use. DMAIC is as you say has more rigor and has a tendancy to require experts to use. Hence the inclusive v exclusive line.
Another thought based on years of observation is that organisations see the tool and think it will solve their problems. My analogy is that when your car is broken you don’t go to your toolbox for the correct spanner until you have understood the problem. What I observe is organisations buying a spanner in the form of Lean or 6 Sigma for example before they fully understand the problem. They then try to make the spanner fit the problem.
My own experience was that my first Green Belt project turned out to be a ‘sledge hammer to crack a nut’ it took 6 months of pain to outcome a result. I could’ve arrived at the same result in a fraction of the time using A3 Thinking.
Both methods have a place but that place must be determined by the problem otherwise I’m wandering around with a spanner trying to find loose bolts that it can tighten.
Regards
Steve.
Wow! Great post and a great white paper by Ron & Cliff! Just one minor comment about Dr. Deming being the “father of the Total Quality Management movement”. As far as I know, Dr. Deming has never liked to use the term TQM and was in fact quite critical to using it to describe a quality management practice which was quite different from his own. Professor Peter B. Petersen writes more about this subject in his article “Total quality management and the Deming approach to quality management” published in the Journal of Management History (1999).
http://www.emeraldinsight.com/journals.htm?articleid=871973
Alex . you are correct . anytime I saw anyone use the term TQM when Dr. Deming was present was in for some questions some teaching . I can hear him now “TQM … what’s that?” and then he would go from there. He wanted to help people think deeply, programs and labels tended to keep that from happening. He was about “transformation of management through understanding and application of the system of profound knowledge” . that wont fit into an acronym or a program . Mike
I’m certainly no Dr Deming, but I like to play the same game with hearing “lean sigma.” What’s that?
Six sigma was all the rage during the later years of my interactions with Dr. Deming, I was at ASQ at the time . If it was brought up in a conversation, you got some of the same reactions as with TQM or the use of PDCA . you also got “the look” . He used to “zing” zero defects during the red bead experiment, I understand he did the same with six sigma a few times (I was not present) He did leave us some clues for us to understand what (I think) was behind this look. Read chapter 10 of New Economics – the very last line is a quote from Don Wheeler “conformance to specifications, zero defects, six sigma quality and other nostrums all miss the point”.
I completely agree that “zero defects” is worthless (harmful) as a slogan or an admonition to the workers (and I’m assuming that’s what Dr. Deming was talking about)… but if something can truly be error proofed (poka yoke), zero defects is possible for a particular process.
I agree the admonition of “zero defects” will not work (the red bead foreman always tried it on day 3, and it always failed) . I think Dr. Deming would ask “by what method” could we improve? and would agree that zero is possible, but it all depends on the method.
Yep, I totally agree. I still think that there is nothing more profound than Dr. Deming’s system of profound knowledge. Sad that it has never been “operationalized” in the same way as lean, 6 sigma or even TQM did.
“Adjust plan” is definitely better than “Act” or “Action” which is often misinterpreted as “action completed”.
In the third part of the cycle, it is important to check both process and results. It is not just about studying the results of an experiment. It is checking for bias in how you took actions, checking your assumptions, checking that the action itself was actually followed through, etc. We are not studying how the process was done, we are simply checking to see if it was done, and whether it was done according to plan. In short, we are checking to make sure that the result is not just due to luck or variation.
We can get there with “study” but we need to be crystal clear in the definition of what is included in that step. The nuance of “study the process” and “check the process” is different, so I prefer check as it is used by Toyota.
[…] and PDSA from Velaction PDSA with Hoshin Kanri from A Lean Journey Prefering PDSA over PDCA from Karen Martin/Lean Blog PDSA Video from IHI at Lean Blog PDCA and Deming Cycle at iSixSigma Problem Solving at the Karen […]
[…] a great guest post by Karen on Mark Graban’s blog about PDSA vs. PDCA, and another spirited post by Mark himself on the […]