6.06.2007

Theory of Constraints: Was Goldratt Wrong?

While I don’t normally focus here on the books we publish, I wanted to discuss one new book because of its intriguing challenge to a long-accepted idea regarding the theory of constraints (TOC).


            As many of you are aware, TOC was pioneered more than 20 years ago by Eli Goldratt and Jeff Cox in the business novel The Goal.


Theory of constraints focuses on maximizing throughput and does so by identifying and eliminating constraints on production, often referred to as bottlenecks.


            (Disclaimer: I am not an expert on TOC, so if you believe I don’t explain it very well, by all means please post a comment on the blog.)


            In TOC, you first identify the system’s constraint(s), then decide how to exploit the constraint(s). Next, subordinate everything else to that decision. Then elevate the constraint. If, in any of the preceding steps, the constraint has been broken, go back to the first step.


            One concept of TOC comes under fire in Beyond the Theory of Constraints: How to Eliminate Variation and Maximize Capacity. This book, which we have just added to our website (copies will be available in August), is written by William Levinson, a manufacturing expert and the author of several other books, including Henry Ford’s Lean Vision: Enduring Principles from the First Ford Motor Plant.


            The TOC concept at issue here has to do with variation. The following description from the Superfactory website explains it:


 


            Another key concept of the Theory of Constraints is that variation (in production and material transfer times) prevents the operation of a balanced factory at 100 percent capacity. This concept is illustrated in Goldratt and Cox's The Goal by a matchsticks-and-dice simulation in which the players represent production stations.


During each turn, each player passes the lesser of his dice roll (his station's capacity for that turn) and the number of matchsticks he has (work waiting at his station) to the next person. Although each station has a theoretical average capacity of 3.5 units per turn, the simulated factory's overall production is somewhat less because high die rolls, which are wasted when no work is available, do not make up for the low ones.


 


But Levinson sees things differently. He contends that much of the variation in processing and material transfer times comes from special or assignable causes that can be eliminated through traditional quality management techniques. He points to the success of Henry Ford, whose product system was designed explicitly to suppress variation in processing and material transfer times.


Levinson believes that the matchsticks-and-dice exercise is useful in teaching the effects of variation on throughput and inventory. However, he contends it may also teach the lesson that the factory is at the mercy of this variation. The purpose of Levinson’s book is to teach the reader how to identify and remove these assignable causes and, as he puts it, “roll a six every time.”


Do you agree? Do you believe it is possible to eliminate these types of variation? Please post your comments, and I’ll hope to revisit this issue after Levinson’s book has been out long enough for people to be familiar with it.


 


 

9 comments:

Ralph Bernstein said...

IMPORTED
6/8/2007 2:05:49 PM
Re: Theory of Constraints: Was Goldratt Wrong?
by: mgraban

Yes, Levinson makes a very good point -- variation shouldn't be taken as a given. Reducing variation should be the goal, not just dealing with it.

Separate point:

One major difference I recall is that Goldratt preached that every line SHOULD have a bottleneck and that, ideally, the bottleneck is first. This way, WIP in a line is limited (the bottleneck limits how much is ever introduced into the process).

Lean teaches you to have a balanced line, no bottlenecks. Lean/TPS also limits WIP in different ways (through Standard WIP, kanban, visual management, etc.).

Ralph Bernstein said...

IMPORTED
6/14/2007 12:18:59 PM
Re: Theory of Constraints: Was Goldratt Wrong?
By: dek

I am looking forward to Mr. Levinson's book. Although this isn't entirely new ground. Hooper, et al, in "Factory Physics" provided a very good explanation of the effects of variability on utilization.

Also, I agree with mgraban -- TOC is based on having an identified and managed bottleneck. To the point where a bottleneck should needs to be 'developed' (not necessarily by slowing down a process) to focus process management. TPS is focused on balanced lines, with no bottleneck.

Ralph Bernstein said...

IMPORTED
6/25/2007 12:35:56 PM
Re: Theory of Constraints: Was Goldratt Wrong?
By: thopper

Goldratt was not wrong, though he may have paid too little attention to special cause variation. In Goldratt's matchsticks-and-dice demonstration, the variation is built into the process, and is therefore normal, not special cause, variation. Such variation will, in all cases, force a trade-off between increasing WIP and accepting lower utilization on nearly all operations.

That said, Levinson's contention that "much of the variation...comes from special or assignable causes" may also be true. If so, the bottleneck should jump from one operation to another at random, and identifying "the" bottleneck will be impossible. Still, identifying bottlenecks, as the Theory of Constraints suggests, should work as well for identifying bottlenecks due to special causes as it does for bottlenecks due to normal variation. It appears that Levinson is admonishing us to not just assume normal variation, but to look for and correct special causes. This is probably good advice.

Ralph Bernstein said...

IMPORTED
6/28/2007 7:56:19 PM
Re: Theory of Constraints: Was Goldratt Wrong?By: rossr

A number of years ago new leadership striving to transition a company from mass to lean emphasized constraint management. The workforce struggled to understand what a constraint was. During a class the VP of Ops was asked what the difference was between Constraint Management and the 7 wastes of TPS. He responded by stating that there wasn't any difference, he just want the associates to focus on the biggest constraint. The problem was that the 7 wastes had never been taught to the workforce to have a basis to identify the biggest constraint. Therefore, Lean or TOC?

Ralph Bernstein said...

IMPORTED
6/30/2007 10:33:27 AM
Re: Theory of Constraints: Was Goldratt Wrong?By: systhinc

I totally agree with Mr. Levinson's viewpoint. Much of the confusion can be taken out of the discussion if we define a constraint as any point in the production system in which waste is encountered. Using Pareto analysis, we eliminate the greatest wastes moving on to the others. Lean, TOC, and traditional TQM are different names for the same ideas.

Anonymous said...

I find it interesting that there is all of this debate...especially given that Toyota is now implementing TOC...seems buffers are a good thing..

SMV said...

I have implemented both TOC and Lean solutions on mfg lines. The key difference between the two is the value placed on elimination of variation and waste.

Lean is very focused on variaton and attempts to eliminate it over time.

TOC achieves a predictable level of flow and then changes its focus on growth. One assumption is that their is limited resources in any organization & that it is better to focus on growth than continued waste reduction.

David said...

The Goal is not concerned with a plant which produces one product on a production line, it assumes a mixture of products utilising some (not all)common production processes. This is the fundamental misunderstood problem associated with lean methodologies from production line plants and why they don't work well in mixed manufacture.

Anonymous said...

My brother recommended I may like this blog. He was once totally right. This submit actually made my day. You can not believe just how a lot time I had spent for this information! Thanks!
banditcustomauto.com |

techhelperusa.com |

northtexasautoplex |

http://www.autoacross.com |

http://www.solotravelapp.com |