Although only 30 years old, I tend to lean my quality styles to the old tried and true methods. I track continual improvement issues under a PDCA format. A management group recently observed this report. They (management) are caught up in the six sigma buzz, requested that I report on issues in the DMAIC format.

What's the fundamental difference? Are both not methods of continual improvement? Or are the deep philosophical differences between the two? Is one more appropriate under a certain set of conditions than the other?

Help me understand the differences. :confused:

I found this - I'll leave the explaining to the experts... In my eyes, they're the same thing with different labels.

http://Elsmar.com/gif/DMAIC.gif

As far as the PDCA, see *** DEAD LINK REMOVED *** for some details.

http://Elsmar.com/gif/PDCA.gif

The concept of the PDCA Cycle was originally developed by Walter Shewhart, the pioneering statistician who developed statistical process control in the Bell Laboratories in the US during the 1930's. It is often referred to as `the Shewhart Cycle'. It was taken up and promoted very effectively from the 1950s on by the famous Quality Management authority, W. Edwards Deming, and is consequently known by many as `the Deming Wheel'.

Plan to improve your operations first by finding out what things are going wrong (that is identify the problems faced), and come up with ideas for solving these problems.

Do changes designed to solve the problems on a small or experimental scale first. This minimises disruption to routine activity while testing whether the changes will work or not.

Check whether the small scale or experimental changes are achieving the desired result or not. Also, continuously Check nominated key activities (regardless of any experimentation going on) to ensure that you know what the quality of the output is at all times to identify any new problems when they crop up.

Act to implement changes on a larger scale if the experiment is successful. This means making the changes a routine part of your activity. Also Act to involve other persons (other departments, suppliers, or customers) affected by the changes and whose cooperation you need to implement them on a larger scale, or those who may simply benefit from what you have learned (you may, of course, already have involved these people in the Do or trial stage).

Maybe an opinion that will be shot down,

6S is nothing more than process improvement. Many companies are investing alot of maney to prove they are 6s when there are already similar processes that have been around for years. One I like is Ford's Q.O.S..

Why spend extra money when we should all be doing this in the first place, process measurables and result measurables provide the data needed for any company to assess their standing and needs.

Now waiting for the fallout, although it won't change my mind.

I have been working with TQM since the late eighties and PDCA has always been at the heart of the philosophy. From this philosophy the company I worked with at the time expanded the PDCA cycle to an 8 step process for problem solving and continuous improvement but never lost it roots from the Deming Cycle. It was defined as:

PLAN
1. Define the Problem
2. Implement/Verify Interim Action¨
3. Aquire and Analyse Data
4. Determine Root Cause(s)
5. Evaluate Possible Solutions

DO
6. Implement

CHECK
7. Verify the Results

ACT
8. Standardise and Future Actions

This approach seems to tie PDCA and DMAIC together quite nicely.

I have been using this system very succesfully for nearly 15 years in 6 different companies and in 3 different European countries.

I am a naturally lazy guy so I am always on the look out for tools that makes my job easier. I have never seen an easier tool that can incorporate all other tools, I have never been shown a better approach and I have never found anyone one to criticise it. Also customers love it as a Corrective Action Report. Just ensure that Step 1 is weakness focused, succinct and does not suggest a cause.

More info available if required.

Alan:thedeal:

Hi!

Since 1998 we are training Six Sigma Green/Black Belts through the PDCA cycle. Definitely there are no differences between them.

Whirlpool of Brazil, Johnsonn & Johnsonn of Brazil, International Motors of Brazil and a great number of firms have their personnel trained with us through the PDCA cycle. All our products are "PDCA based", including "Six Sigma Black Belts".

Perhaps a little more time is needed for the business world to fully understand that there are no differences between them.


Cristián Carranza
Senior Consultant
Fundação de Desenvolvimento Gerencial - Brazil

Just wondering if anyone has anything to add to this...

PDCA I understand...it is a well-established methodology within my organization.

DMAIC I know little about, but looking at what it stands for, I'm seeing the following associations:

Define = Plan
Measure = Do
Analyze = Check
Improve = Act
Control = ????

I have said before that there can be no improvement without standardization. Is that what the Control equates to, then?...an established process/practice?

The Control part is institutionalizing the change. In my book, this is the Act part. Once you have Checked and made sure that the Do that you Planned is an actual improvement, then you Act to make the change standard and controlled.

In Six Sigma methodology, this includes a continuing analysis of the Change via Dashboard metrics, etc. to make sure that the change stays stable.

Paraphrased from Six Sigma for Managers, Greg Blue
Consider Y = some function of X
Define = Identify important problems in your processes.
Measure = Select Critical to Quality (CTQ) characteristics in the product or process, define performance standards, validate measurement system, establish process capability (Y)
Analyze = Define improvement objectives, identify variation sources, screen potential causes for change (Y) and identify vital fiew (X)
Improve = Discover variable relationships among vital few, extablish operating tolerances, validate measurement system (X)
Control = Determine ability to control vital few, implement process control system on vital few (X)

Just wondering if anyone has anything to add to this...

I should point out that Dr. Deming and others got away from "Check" and use "Study" instead. Thus, PDSA cycle. This change is what differentiates DMAIC vs. PDSA. In "study" we study the results, non-judgementally. We see if an improvement occurred, preferably using SPC. We should NOT compare against some numerical goal that was established back in the Measure stage, as the six sigma folks would tend to do. Then we do the final "Act"

I think the weakness of DMAIC is that final "C" for control. That implies to me more of an endpoint to the process. PDSA is definitely intended to move from the Act stage back to Plan to see if anything further can be done on the next cycle.

Another difference is that the "Do" in PDSA is a small scale prototype. DMAIC just moves straight into "implementation" after measuring and analyzing.

Great stuff!!!

Nothing in Quality is ever new. I wrote my first ISO 9001:1994 manual using a 1962 Quality Manual for refrence.

Great info Steve!

I still have to go back to my original view post. Much of what is going on is semantics while the underlying tool is still the same. But as I have read a few threads on this, I do agree (albiet grudgingly:mg: ) that as long as the job gets done, who cares what name we use.

"A=A", even if we call it "X"

Al...

Great stuff!!!
Nothing in Quality is ever new. I wrote my first ISO 9001:1994 manual using a 1962 Quality Manual for refrence.

From The New State, by Mary Parker Follett, 1918:

"We sit around the council table not blank pages but made up of all our past experiences. Then we evolve a so-called common will, then we take it into the concrete world to see if it will work. In so far as it does work, it proves itself; in so far as it does not, it generates the necessary idea to make it 'common.' Then again we test and so on and so on. In our work always new and necessary modifications arise which again in actualizing themselves, again modify themselves."

If anyone knows of a solid, published reference to PDCA before this one, please let me know.

This is where my "TS16949 Awareness" introduction presentation begins.

Nice.. Glad you agree!

I should point out that Dr. Deming and others got away from "Check" and use "Study" instead. Thus, PDSA cycle. This change is what differentiates DMAIC vs. PDSA. In "study" we study the results, non-judgementally. We see if an improvement occurred, preferably using SPC. We should NOT compare against some numerical goal that was established back in the Measure stage, as the six sigma folks would tend to do. Then we do the final "Act"

I think the weakness of DMAIC is that final "C" for control. That implies to me more of an endpoint to the process. PDSA is definitely intended to move from the Act stage back to Plan to see if anything further can be done on the next cycle.

Another difference is that the "Do" in PDSA is a small scale prototype. DMAIC just moves straight into "implementation" after measuring and analyzing.I was vaguely aware of "study" versus "check" before today. However, the most important point for me is the "cycle" part. The circular diagram which is the traditional way of displaying PDCA is the foundation of continuous improvement and every Design of Experiment. It is a fundamental difference in approach which is sort of "short-circuited" by the "control" aspect of DMAIC being perceived as a final, stable process versus a springboard for a new "design" or "plan."

Hi all!

With regard to this slight difference between PDCA and DMAIC, I wish to add some additional comment:

PDCA means "improvement" and SDCA (Standard-Do-Check-Act) means "sustain de change" (daily routine work management, the way the Standard Operational Procedures are followed).

According to this concept, the Control step in DMAIC methodology is clearly associated with the SDCA cycle.

It is not the same thing to improve a process and to sustain the change...

Regards.

Good point - A definite difference twixt the two.

Hi all!

With regard to this slight difference between PDCA and DMAIC, I wish to add some additional comment:

PDCA means "improvement" and SDCA (Standard-Do-Check-Act) means "sustain de change" (daily routine work management, the way the Standard Operational Procedures are followed).

According to this concept, the Control step in DMAIC methodology is clearly associated with the SDCA cycle.

It is not the same thing to improve a process and to sustain the change...

Regards.

I agree. :agree1: It would appear to me, then, that the "I" and "C" in DMAIC are at the same level with with a possible 'flow' going from "C" to "I", as you can only improve once everything is under control.

I agree. :agree1: It would appear to me, then, that the "I" and "C" in DMAIC are at the same level with with a possible 'flow' going from "C" to "I", as you can only improve once everything is under control.

And how many times have you come up with the perfect process at the first attempt to improve it? :nope: Me, I'll stick with PDSA cycles and continually improve. :yes:

And how many times have you come up with the perfect process at the first attempt to improve it? :nope: Me, I'll stick with PDSA cycles and continually improve. :yes:

We're good...not perfect. :notangel:

That's why once we Standardize a process (SDCA), only then will we attempt to improve it (PDCA). The C in PDCA is Check. A (or Act) can result in one of two things. If we got it perfect on our first attempt :agree1: then we Standardize the process (SDCA) at a new level of performance. If the expected results were not achieved, the A is to go back to the drawing board, do some more analysis, and come up with a new game plan. :read:

Excellent, Roxane. You are truly a Deming Disciple!
The good Doctor was constantly warning against attempts to change or adjust processes that were not stable or not understood, usually resulting in making the problem worse.

We're good...not perfect. :notangel:

That's why once we Standardize a process (SDCA), only then will we attempt to improve it (PDCA). The C in PDCA is Check. A (or Act) can result in one of two things. If we got it perfect on our first attempt :agree1: then we Standardize the process (SDCA) at a new level of performance. If the expected results were not achieved, the A is to go back to the drawing board, do some more analysis, and come up with a new game plan. :read:
Yes, I will agree that when we first go into a process we need to understand what it has been doing, and whether or not it is statistically in control. No argument there. I will suggest that this is what the first pass through PDSA should do - during Plan you gather data and make the control chart to see if it is in control.

I am still wary that the phrase "if the expected results were not achieved" could turn into Management by Objective - comparison to an arbitrary target. But if the phrase is used in the context of - we did not see on the control chart a change (either the process coming "into control" when it had not previously been, or an improvement in the baseline average if it was previously in control) then I am in agreement.

Excellent, Roxane. You are truly a Deming Disciple!
The good Doctor was constantly warning against attempts to change or adjust processes that were not stable or not understood, usually resulting in making the problem worse.

Thanks. :o To be honest, though, I have never read any of his books...but someone in our parent company must have, as the philosophy of "There can be no improvement, without standardization" has been drilled into us. Some of our newest family members have experienced some growing pains with this approach, but they are now beginning to see the benefits to first understanding where you are and identifying process capability before focusing on improvement projects.

I am still wary that the phrase "if the expected results were not achieved" could turn into Management by Objective - comparison to an arbitrary target. But if the phrase is used in the context of - we did not see on the control chart a change (either the process coming "into control" when it had not previously been, or an improvement in the baseline average if it was previously in control) then I am in agreement.

I will admit that on our improvement projects we do have goals/objectives...all based on analysis of the data and the theoretical planning of the results from our actions. If we do not achieve the goal/objective, we analyze why...and if the reason is logical and supported by evidence, we accept the results. But to not have goals is unacceptable to us...to simply say "We improved on our energy consumption last month" is meaningless to us. We want to see how we did based on what we did. Did our actions contribute as much as we thought they would? If not, why? Not only does this help us on our current project, but this will also provide us with planning knowledge and awareness on future improvement projects.

I thought I pick this interesting thread up again.

Earlier in this thread the discussion was about the circularity of PDSA and I want to call it linearity of DMAIC, meaning there is no provison for more iterations.
Wouldn't that be the case, too, for PDSA if Shewart or Deming wouldn't have drawn it as a circle (Deming Wheel)?

PDSA alone does not imply to me that I immediately start looking into the same process again. And I think that that is not possible in real live anyways since there are more processes that need improvement. Once other processes are adressed I will automatically return to a process that was adressed before, since now it's back on top of the list.

The same thing for DMAIC. If projects are chosen the right way eventually every process needs to be adressed again. So I would say DMAIC is also a circle (as drawn by Marc in the second post) but between C and D is a time delay as is between A and P in the PDSA cylce.

I thought I pick this interesting thread up again.

Earlier in this thread the discussion was about the circularity of PDSA and I want to call it linearity of DMAIC, meaning there is no provison for more iterations.
Wouldn't that be the case, too, for PDSA if Shewart or Deming wouldn't have drawn it as a circle (Deming Wheel)?

PDSA alone does not imply to me that I immediately start looking into the same process again. And I think that that is not possible in real live anyways since there are more processes that need improvement. Once other processes are adressed I will automatically return to a process that was adressed before, since now it's back on top of the list.

The same thing for DMAIC. If projects are chosen the right way eventually every process needs to be adressed again. So I would say DMAIC is also a circle (as drawn by Marc in the second post) but between C and D is a time delay as is between A and P in the PDSA cylce.

I am agree with "If projects are chosen the right way eventually every process needs to be adressed again.". :agree1:

From my point of view PDCA, SDCA, PDSA, DMAIC, and some other version that anyone of us have developed to improve processes are closely related eachother, just because they all are coming from the same principle. Through the last 15 years or more we have learned that process improvement are more effective (from business stand point) if we analyse it as a system (Dynamics system). Whenever we develop that sense into our organization our inicitaives become more effective or have greater impact on business results.

So what I am trying to say is that analyse processes without understand what is happening upstream and downstream can took away our focus from which is real important: Improve business results.

I think that the first step is understand the interactions between processes and them we can priorize which process or processes should be attend it first to improve business performance. To improve processes we should use the method that our organizations have defined as a way to do it (PDCA, DMAIC or any other version).

There are two inplicit steps that I do bring out to be more specific and make everyone involved with business results.

1.- Define the data sources and verify the fidelity of the data given. This is before or part to the "P" or "D" stage (Depend which method we choose).
2.- Deploy lesson learned through the operational crew to keep them commited with business results and keep Know How alive. This must be added to any of both method just because none of them make a call for people education (operational crew).

HAVE A NICE WEEKEND TO EVERYONE!. :cool:

Guys....and Girls all these tools work.....it just depends on the title it was given at your last school our course. As ISO9001 is catching up we can continue to lead them in improvement strategies. Different techniques or the same technique by a different name works differently with different work groups. Don't stress yourselvs about it.....just do it!