Greetings all....

I thought I might start off this topic with a conundrum that I ran into today.

I just received the September 2001 edition of Quality Digest, in which Thomas Pyzdek, in his article "Six Sigma: Quality's Evolution?", tabulates the differences between Six Sigma vs. Quality. One of the differences listed under the Six Sigma column, he states, "Goals flow down from customers and senior leadership's strategic objectives. Goals and metrics are review at the entrerprise level to assure that local suboptimization does not occur."

Now, according to Dr. Deming, "The obligation of any component is to contribute its best to the system, not to maximize its own production, profit, or sales nor any other competitive measure. Some components may operate at a loss to themselves in order to optimize the whole system, including the components that take a loss." {The New Economics, page 100}.

My take on this is that Dr. Deming is correct in his ascertion, and I'll illustrate with an example:

Process "A" manufactures widgets that feed into Process "B" which transforms those widgets into thingama-jigs. The output capacity of Process "A" is 1000 widgets per hour. The output capacity of Process "B" is 750 thingama-jigs per hour. The capacity of Process "B" cannot be increased without significant capital expenditure. Market research has indicated that there is sufficient demand for thingama-jigs to warrent that expenditure, but the capacity of Process "B" can only be increased to 900 thingama-jigs per hour.

The money is spent on new equipment and Process "B" is now increased to 900 thingama-jigs per hour. Now if Process "A" is optimized, every hour there are 100 excess widgets sitting on the shop floor waiting to be moved into Process "B". At the end of an 8 hour shift, there will be 800 widgets in "work in process" inventory, or at the end of the week, there will be 4,000 widgets in "work in process" inventory. The cost of the widgets is now a "sunk cost" which ties up money which might be used elsewhere.

So, what benefit is derived from assuring that local suboptimization of Process "A" is not occuring?

Any thoughts, comments, critiques?

Cheers!!!

Mike

The benefit of not suboptimizing Process A is that you do not build unwanted inventory, create flow, and work efficiently. Unwanted inventory is waste that the Customer does not want to pay for, but does when organizations ignore this lesson.

Brian Joiner uses a clever example of a team of rowers. If each were to do his individual best (maximize effort), the boat would splash around and go nowhere. However, if each rower were to work in harmony, the scull would effortlessly glide across the river (optimization of a system). Some would obviously not work as hard as others, but the net result of everyones efforts would be more than any one persons achievement (synergy).

Dr. Deming's example of a Cafeteria running for a loss to the benefit of the organization is a classic example of how one component might sacrifice itself for the whole.

Regards,


Kevin

Hi Kevin,

I completely agree.... perhaps I worded my question poorly. I am wondering why a firm would want to "assure that local suboptimization does not occur". It seems to me that Tom Pyzdek is espousing optimization of all processes and subprocesses regardless of the impact on throughput, downline process capability and inventory levels. This just doesn't seem logical to me.

Any thoughts?

Cheers!!

Mike

Hello Michael,

I like the Senge quote by the way.

As I am sure you are aware, Tom is a Statistician and has written a book or two on SPC before jumping on the Six Sigma bandwagon. As I have not read any of his books, my comments may be way off base, but here they are just the same.

As a proponent to SPC, he may be promoting the use of process controls without regard to suboptimization. In addition, the Six Sigma wave is primarily focused the pieces rather than the whole. Suboptimization appears to be the key to financial results. Regardless of what later Six Sigma authors have written, the comments directly from Mikel Harry clearly show that short term success through suboptimization is the factor to better profits and dividends. He is right if you don't look beyond the immediate results. However, he is dead wrong about creating constancy of purpose.

Mikel Harry views the Shareholder and the Organization preceeding the Customer. As such, the recipe for success in his program is suboptimization. Later authors have tried to close the gap in his program by adding emphasis to the Customer Satisfaction angle. Some have even tried to blend Dr. Deming's SoPK with Six Sigma. I'm not sure where Tom stands on this point, but I was abit disappointed when he jumped on the SS wagon.

Goldratt offers his three key business measures, amongst them he lists Inventory and Throughput. While I disagree with him on many aspects, I believe that these two measures are important to any organization. As such, they need to be measured and improved but as Goldratt states, they must be done in regard to the other two measures (he includes Operating Expenses). Although his view is a limited System View (mostly relationships), his point is valid. Optimization of the pieces leads to suboptimization of the whole. If Tom Pyzdek is suggesting the opposite, I fear he is also dead wrong.

Regards,

Kevin

Originally posted by Kevin Mader
Hello Michael,

I like the Senge quote by the way.



Thanks Kevin - I'm a big fan of Senge - The Fifth Discipline is one of my all time favorites.

Unfortunately, I believe you are correct with regards to 6S being focused primarily on the pieces rather than the whole. In addition (from what I've read), it focuses on a dollar figure too. I find that disturbing since some critical changes necessary for customer satisfaction, process improvement, etc., may not have a big $$$ return and would be turned down as a 6S project. I think that aspect also fails to recognize the potential loss that can be incurred should a project not be undertaken. But then, how can that be measured? If it can't be measured, 6S doesn't want to deal with it.

Goldratt is good at focusing on processes (as well as telling a compelling story), but falls short when incorporating the quality aspect of process improvement, etc. I've only had time to read The Goal, so am not familiar with his other works.

Thanks for your input!!

Cheers!!!

Mike

Hello Michael.

Your observations of Six Sigma are very similar to my own. How do you measure a satified customer? How about a dissatisfied one? As I am sure you are aware of Lloyd S. Nelson's comment, "Most of the important figures are unknown and unknowable."

In addition, Dr. Deming stated in a letter to Dr. Joiner (foreword): "...It is wrong to suppose that if you can't measure it, you can't manage it. Most of the heavy losses caused by management today cannot be measured, yet they must be managed. Thus, management may spend $20,000 to train six people in a skill. This $20,000 is an investment. Its magnitude is known, $20,000. The future benefit (return on investment), however, will never be known, cannot be measured. Management's action is based on theory, prediction that the investment will pay off handsomely...", page ix, Fourth Generation Management.

Those organizations using only 'hard numbers' to determine projects are missing most of the important factors to doing business.

Regards,

Kevin

Hi Kevin,

While I didn't intend this thread to be 6S bashing, and there are quite a few good things about 6S, the thing I find most distressing (as I've stated previously) is the use of dollar figures to determine whether the project is worth doing.

You are absolutely right, solely relying on hard numbers to make business decisions is dangerous. There are just too many factors that cannot be quantified (customer dissatisfaction is just one of them) to base business decisions solely on numbers. Unfortunately, colleges & universities are still teaching that type of management, existing upper management is using these types of criteria to determine who gets promoted (who is the rising star), etc. And still -- no one is paying attention to Mr. or Ms. Customer. Sure, a great many businesses say they do, but how many of them truly practice what they preach. For example, I attend a monthly "Network" meeting hosted by one of the community colleges that discusses things like Tools & Methods of Quality, etc. I recently brought up the topic of using QFD and applying it to help satisfy some of the Customer Focus issues in the new ISO 9000:2000 standard. Only 2 or 3 people (these people are quality managers, VP's, etc.) had even heard of QFD. So, the end of this month I'm giving a workshop on an introduction to QFD... we'll see how it goes... ;)

Sorry for the beginnings of a rant... :D

Cheers!!

Mike

Mike,

Perhaps this should be in a thread of its own, but I fear that it would only create a critical and negative discussion. I will place it here for open and constructive debate, especially along the lines of this threads topic, optimization vs. suboptimization.

Looking at Dr. Deming’s System of Profound Knowledge, we know it is comprised of four key components: Appreciation of a System, Theory of Knowledge, Theory of Variation, and Theory of Psychology. These are guiding principles. They work together and are equally important. Honoring one without consideration to the other is suboptimization. I have my favorites for sure, but I believe this to be true.

Now Dr. Harry’s Six Sigma program operates differently. It has many of the same Quality tools, but what are the guiding principles? A philosophy must have underlying guiding principles to in fact be – a philosophy. Yet, I never see them discussed in the many papers, posts, and articles written on Six Sigma that I have read. Perhaps it is in one of Dr. Harry’s books (I have not read any books on Six Sigma although I have browsed many Tables of Contents). Still, I find it curious that I can find no evidence of any underlying guiding principles. This is why I cannot endorse Six Sigma as a philosophy as many claim it to be. By their definitions, they are right, I suppose. By mine though, they are wrong. Six Sigma is devoid of Profound Knowledge even if it shares many of the Quality tools found in other programs I feel have merit. Tools are not philosophies.

My filters are tuned to the SoPK, so perhaps my viewpoint is heavily skewed. When considering aspects of Six Sigma using these filters, I find it to be immoral. The fact that the program is ‘results-oriented’ instead of being ‘process-oriented’ despite its heavy dependence on Statistical Process Control. The fact that it uses rewards, incentives, and other Extrinsic motivators (if such a thing exists) in place of intrinsic motivation, pride and joy in work. The fact that it is short-term goal oriented that best have a payoff worthy of a CFOs approval (defect reduction for the Customer is an ancillary benefit, IMHO). The fact that it places the needs of the ‘Organization’ and the ‘Shareholder’ (a supplier) above the needs of the Customer. For me, this is pure rubbish and compelling evidence why Six Sigma only promotes the Western Management Philosophy and prolongs this dieing paradigm. It will only worsen the state we are in!!

I think it is fair to state the obvious oversights of the Six Sigma program. While some might see this as bashing, I would argue that it is a matter of perspective. Hearing that you have adopted a plan that is based on questionable practices would upset anyone. Imagine all those who have given Performance Appraisals to their subordinates for decades hearing today that “Performance Appraisals are destructive”. It is difficult to admit that we were wrong for so long, almost unfathomable. Those who do not challenge their assumptions will often fall victim to them.

Some plans are better than others. There are good points to most any plan. What are they in Six Sigma? Are they unique to Six Sigma, or were they part of some other plan sometime earlier? While Six Sigma may have many good points, my issues with the program deal on the human level. What level of Psychology does Six Sigma practice? The carrot and the stick approach to suboptimization, I’m afraid.

Well, enough of my rant.

Back to the group…

Kevin

Contemporary comments?

It seems that the question asked by the OP is misinterpreted, because the word "optimize" is misdefined (as is "synergy" in Kevin's first response). To optimize a process is to make it as good as it can be under the circumstances. So a "suboptimal" process isn't working as well as it's capable of working, given known (and unavoidable) constraints.

The concept of "synergy" involves two or more components of a system whose output or performance is greater than the individual components' output independently summed. It's a term from biology, and doesn't fit well in other contexts.

Whether or not a component process is "suboptimal" takes into consideration what's expected of the whole system. So if operation A, operating independently, is capable of producing 1000 units per hour, and operation B is capable of 750, slowing "A" down to 750 isn't evidence of suboptimization. On the contrary, if the best possible combined output is 750 per hour, speeding A up to its capacity would result in "suboptimization."

Jim,

Processes and sub-systems can be suboptimized to a higher or lower level than what is optimal. In order to optimize the system, some components will be suboptimized. Some examples were listed above as well as your own contribution. There are also non-biological uses of the word ‘synergy’ which can be defined as ‘working together’.

What I can’t be certain of, given your comments and through a review of Mike’s original post, is which type of suboptimization Tom is speaking about. It does appear that he is speaking about localized suboptimization that in my experience generally means maximization of a process or subsystem to the detriment of the whole. Tom may instead be warning the reader not to create localized suboptimization of this type as by maximizing an area or process will negatively impact the system optimization as a whole. Six Sigma, in my experience, attempts to create localized optima. This often times is suboptimal to the System.

My assumption, right or wrong, was that Tom’s usage of suboptimization was an endorsement to the creation of localized optima.

Regards,

Kevin

P.S. I have still not read Tom’s article. In fairness to him, the discussion here should probably center about Mike’s selection of this threads header: optimization vs. suboptimization. Oddly enough, it’s really not a competition.

Processes and sub-systems can be suboptimized to a higher or lower level than what is optimal. In order to optimize the system, some components will be suboptimized. Some examples were listed above as well as your own contribution.
Sigh. It's never a good idea to redefine words in order to get them to mean what you want them to mean. "Optimum" refers to the most favorable conditions for something or, as I said earlier, making something as good as it can be under the circumstances. The "big" system can't be optimized if any of its components aren't optimized. Do the math. "Optimum" is not necessarily related to "capacity."

There are also non-biological uses of the word ‘synergy’ which can be defined as ‘working together’.
I said the word comes from biology, but I didn't say there were no legitimate uses outside of biology. This isn't one of them, though. It refers to a whole that is greater than the sum of its parts, not just "working together." "Symbiosis" might have been a better fit, but it's not right either. Why not just say "working together"?

"Optimum" refers to the most favorable conditions for something or, as I said earlier, making something as good as it can be under the circumstances. The "big" system can't be optimized if any of its components aren't optimized. Do the math...

No, I think Kevin is right. We are not using words in a sterile or clinical environment. We are trying to optimize processes, not describe process already in an optimal state.

The process of doing the optimizing has lots of situations where we are not yet in a favorable state.

For example, we sweat and strive and stress, argue with clients and coworkers, fight for resources, careers start and end, compromises are made...in order to come as close to optimizing (not reaching optimal).

In that context, a non-teamplayer, could carve out a little special situation for himself, to the detriment of the whole. In industry, we refer to that as sub-optimizing.

No, I think Kevin is right. We are not using words in a sterile or clinical environment. We are trying to optimize processes, not describe process already in an optimal state.

The process of doing the optimizing has lots of situations where we are not yet in a favorable state.

For example, we sweat and strive and stress, argue with clients and coworkers, fight for resources, careers start and end, compromises are made...in order to come as close to optimizing (not reaching optimal).

In that context, a non-teamplayer, could carve out a little special situation for himself, to the detriment of the whole. In industry, we refer to that as sub-optimizing.

I think you're misunderstanding my objection, Helmut. In this context, "suboptimization" refers to deliberately causing a sub-operation to perform at something less than its capacity in order to balance its output with the needs of other processes, and ultimately, the uber-process that results in the finished product. Of course, that's just prudent planning. There's no sense in having a lot of WIP lying around all over the place if it can be avoided. My point is that using the correct definition of "optimize," a sub-process is not sub-optimal just because it's not producing at its full capacity. In fact, the opposite is true, in view of the big picture.

I think you're misunderstanding my objection, Helmut. In this context, "suboptimization" refers to deliberately causing a sub-operation to perform at something less than its capacity in order to balance its output with the needs of other processes, and ultimately, the uber-process that results in the finished product. Of course, that's just prudent planning. There's no sense in having a lot of WIP lying around all over the place if it can be avoided. My point is that using the correct definition of "optimize," a sub-process is not sub-optimal just because it's not producing at its full capacity. In fact, the opposite is true, in view of the big picture.


No disagreement there. Perhaps, when used as a verb, the term takes a different connotation.

The usage I usually hear for "suboptimizing" as a verb refers to focusing on benefitting a subprocess to the detriment of the whole. Don't know what Websters says about it, but this is the common usage I frequently hear.

But, I agree with your usage as well.

:topic: I'll admit I didn't think this would digress into a semantics discussion.

:topic: I'll admit I didn't think this would digress into a semantics discussion.


Sorry 'bout that. But, it's an important topic, so making sure we understand the what the topic means seemed relevant.

Yeah - Not a big deal to me. Just a comment. I guess when these forums were smaller there tended to be few semantic issues.

Please view Tom Pyzdek's article on considering constraints qualitydigest.com/june00/html/sixsigma.html

Please view Tom Pyzdek's article on considering constraints qualitydigest.com/june00/html/sixsigma.html

Hello TucsonTom,

Are you referring to this article? http://www.qualitydigest.com/june00/html/sixsigma.html

Stijloor.

I'm sure that's the article TucsonTom wanted to link to. Remember, one must have 8 posts here in the forum before they can put in a live link. This restriction is to minimize forum spammers.

You are both correct.

A rather interesting resurrected thread.

With respect to optimization and suboptimization, if this is in a "systems thinking" context, suboptimization usually refers to "improving" a component of a system, while leading to no improvement or even a degradation of the overall system. Dr. Russ Ackoff uses the example of attempting to put a Rolls-Royce engine in a Geo Metro in order to get a "better" car. The Rolls-Royce engine likely won't even fit within the hood of the Metro, and we would not even end up with an operable automobile, let alone an "optimal" automobile.

The TOC article is interesting. The example given in the diagram is rather simplistic in that it can be solved with a linear program. Here we have only one objective function (maximize profit) subject to a set of constraints on the time available from the four workers. It appears from the later writeup that worker "A" is "slack" - worker A has idle time that is not used in the optimal solution. Thus, any actions spent to improve worker "A" won't improve the output, and would reduce profit if you had spent any money on the effort.

Unfortunately most of our real-life systems have multiple (and competing) objective functions, and there is variability, scrap and other issues. So, the concept of suboptimization becomes even more important, and even linear programming solutions can be woefully inadequate.

...

Unfortunately most of our real-life systems have multiple (and competing) objective functions, and there is variability, scrap and other issues. So, the concept of suboptimization becomes even more important, and even linear programming solutions can be woefully inadequate.


I may have an interesting real world example of the foolishness of "Sub-optimization." I think accountants and controllers should give advice to management teams, but should not be permitted to ENFORCE policies. They usually have too narrow a perspective.

There seems to be a trend developing out of Detroit based automotive suppliers. Several big companies have initiated a cost savings project aimed at reducing printing and ink costs by eliminating small, personal printers. Seems the "bean-counters" have determined their divisions are spending too much money on ink and toner.

Now, we all know that the small printers have a higher per page cost than big, central printers. Unit caost is higher, but there are other costs to consider as well. Productivity, time, confidentiality, data analysis benefits from using color wisely in presentations.

They are banning ALL small, personal printers. EVERYONE must use the central printing stations. All little printers must be discarded, given away, whatever. Even if they are almost new.

The suboptimization results from misguided thinking. There will certainly be savings on ink and toner. That line item on the expense ledger will show great improvement and some bean-counter will get a bonus.

In the meantime, plant managers will lose points off productivity from the greater activities. However, these real costs won't be correlated back to the ink and toner savings.

The ink costs will have been suboptimized, but the greater optimization of the whole production process will suffer. The net cost will be a negative savings. It amazes me that companies are still capable of such foolish thinking, after so much has been studied.

I may have an interesting real world example of the foolishness of "Sub-optimization." I think accountants and controllers should give advice to management teams, but should not be permitted to ENFORCE policies. They usually have too narrow a perspective.

There seems to be a trend developing out of Detroit based automotive suppliers. Several big companies have initiated a cost savings project aimed at reducing printing and ink costs by eliminating small, personal printers. Seems the "bean-counters" have determined their divisions are spending too much money on ink and toner.

Now, we all know that the small printers have a higher per page cost than big, central printers. Unit caost is higher, but there are other costs to consider as well. Productivity, time, confidentiality, data analysis benefits from using color wisely in presentations.

They are banning ALL small, personal printers. EVERYONE must use the central printing stations. All little printers must be discarded, given away, whatever. Even if they are almost new.

The suboptimization results from misguided thinking. There will certainly be savings on ink and toner. That line item on the expense ledger will show great improvement and some bean-counter will get a bonus.

In the meantime, plant managers will lose points off productivity from the greater activities. However, these real costs won't be correlated back to the ink and toner savings.

The ink costs will have been suboptimized, but the greater optimization of the whole production process will suffer. The net cost will be a negative savings. It amazes me that companies are still capable of such foolish thinking, after so much has been studied.

This phenomenon has a written history that dates at least to 1607, and is called "penny wise and pound foolish."

In the meantime, Japanese competitors are eating their lunch by producing better products and providing better service. I worked with the auto industry and Dr. Deming in the early to mid 1980s and watched them improve from 3 sigma to 4 in a very short time, which returned them to profitability. Then they declared victory and progress slowed to a crawl, then stopped completely. Very frustrating! What is it that creates such blind spots as bean-counter management? I suspect it's our business schools, but I have no data to support this suspicion.

I suspect it's our business schools, but I have no data to support this suspicion.

Having taught business evening courses (including MBA courses), YES, our business schools share a considerable blame in this question.

Having taught business evening courses (including MBA courses), YES, our business schools share a considerable blame in this question.

I don't think that business schools teach people to be greedy. There's a fairly complex set of influences at work, including the tendency of the markets to worship short-term gains. While the business schools, and higher education in general, are at fault for graduating people who think that using words like "suboptimization" is a viable substitute for thinking, people entering the business world are most likely to swim with the current, regardless of what they were taught or how well it was taught.

I don't think that business schools teach people to be greedy. There's a fairly complex set of influences at work, including the tendency of the markets to worship short-term gains. While the business schools, and higher education in general, are at fault for graduating people who think that using words like "suboptimization" is a viable substitute for thinking, people entering the business world are most likely to swim with the current, regardless of what they were taught or how well it was taught.
I agree, it isn't necessarily the "greed". What I saw in the textbooks was a completely distorted version of Dr. Deming's 14 points (completely unrecognizable versus the original), and very simplistic data analysis using moving averages. Of course, accounting itself leads to local suboptimization, and also an issue Dr. Ackoff has addressed - business practices tend to punish errors of commission, but completely overlook errors of ommission (not doing something, not spending the money, not making profit from not doing something doesn't show up on the books).

Also, I never saw such a bunch of competitive people as MBA students who had a hard time coopertating on projects. Kind of like (though not so extreme) the behaviours seen on The Apprentice.

I agree, it isn't necessarily the "greed". What I saw in the textbooks was a completely distorted version of Dr. Deming's 14 points (completely unrecognizable versus the original), and very simplistic data analysis using moving averages. Of course, accounting itself leads to local suboptimization, and also an issue Dr. Ackoff has addressed - business practices tend to punish errors of commission, but completely overlook errors of ommission (not doing something, not spending the money, not making profit from not doing something doesn't show up on the books).Exposure to Deming is a good first step, and I'm glad to know that he's even mentioned in MBA classes. Of course we have to hope that the textbooks get it right, but anyone with experience with textbooks knows that they often don't, which is where the expertise of the instructor comes in. Your students were fortunate.

Also, I never saw such a bunch of competitive people as MBA students who had a hard time coopertating on projects. Kind of like (though not so extreme) the behaviours seen on The Apprentice.
The tentacles of greed are not limited to grasping for money. Desire for power and influence is also a strong motivator, and one can be just as greedy for it as for anything else. Hence the phenomena you witnessed, I suspect. There are also psychological influences at work here. People who have yen for power are often confused for those who want to practice leadership. Get a bunch of the former category together, try to get them to work as a team, and you'll quickly see the difference between leadership and greed.

In the meantime, Japanese competitors are eating their lunch by producing better products and providing better service. I worked with the auto industry and Dr. Deming in the early to mid 1980s and watched them improve from 3 sigma to 4 in a very short time, which returned them to profitability. Then they declared victory and progress slowed to a crawl, then stopped completely. Very frustrating! What is it that creates such blind spots as bean-counter management? I suspect it's our business schools, but I have no data to support this suspicion.

Lack of Leadership.

Unwillingness to take risks.

Trying to blame everyone else, no themselves.

Trying to save their way out of this morass.

Taking 25 years too long before they began to actually listen to their customers...