When you are a control engineer in a plant, you tend to do what your boss tells you to do. That’s usually plenty to worry about. The task at hand is to make sure that machines in your area keep running, and to perform occasional upgrades. Your focus is on one machine or part of a line at a time.
Kevin Kohls, now director of Throughput Analysis and Simulation at General Motors Vehicle Operations, in Warren, Mich., found himself in just such a situation as a beginning engineer in a power train manufacturing plant. He heard about throughput being an issue at the plant, but it wasn’t his issue. Even though the “Japanese invasion” was in full swing at the time and survival of the plant was critical, engineers just tended to work on their own areas.
The concept of throughput is rather simple. If the plant doesn’t “make rate,” that is, ship enough product, then it doesn’t make money. Throughput analysis looks at the entire plant, not just parts, to assure that the required amount of quality product is shipped every day. Someone must look at the plant as a whole rather than focus on one machine at a time.
Kohls left GM for a year-and-a-half to earn a masters degree in engineering at Purdue University. Upon returning to the plant, he discovered that it was still not making production. So he started asking questions outside his area. What are the problems that cause throughput of the factory to be less than the design rate? How can they be discovered? Then, how can they be solved?
The goal
Enter GM Research. Programmers there had developed a software tool for predicting throughput and locating bottlenecks in the process. Many people throughout the corporation were suspicious of its utility, but Detroit/Hamtramck Plant Manager Larry Tibbets heard a presentation and liked what he heard. Kohls became part of Tibbets’ implementation team. At the urging of one of the software developers, he read Eliyahu Goldratt’s book, The Goal, and took a Theory of Constraints (TOC) class from the author himself at the Goldratt Institute.
Goldratt’s book explains the TOC through the story of a fictional plant manager who is faced with the imminent closing of his plant because it just can’t manufacture enough products to make a profit. As the plant manager ponders his situation, he meets a Yoda-figure in the person of an older, distinguished consultant. The consultant gently steers him to discover each of the bottlenecks in the plant. As the management team in the book discovers each bottleneck and solves it, they see overall production going up, giving them increased hope of saving the plant. At each step of the way, the consultant would ask the plant manager, “What’s your goal?”
To greatly simplify, the TOC holds that a plant will not produce more than its slowest process or machine. To dramatize the point about organizing processes, Goldratt has the plant manager go along on a Boy Scout outing and hike. Like in most groups of young boys, there are some who are quite athletic and fast, and others who are slower. The Scout leader organized things in the way that seemed most logical to achieve the goal of reaching the end of the hike in the required amount of time. So, he put the fastest boys in front and the slowest ones at the rear.
What happened? The fastest ones ran down the path, then had to stop and wait for the others to catch up. This created a frustrating situation for all. The plant manager, pondering his lessons about throughput, figured out that it would be better if the slower boys were the leaders. It increased their confidence and pleasure, and the whole troop could go at a steady pace, which was fast enough to reach its goal. This is like many manufacturing processes in which one machine is dependent on the output of another machine. There is a lot of “hurry up and wait,” with resulting underutilized resources, as well as scheduling and work-in-process problems.
TOC applies the beliefs, definitions and methods used by the hard sciences to understand and manage the material world to human-based systems (individuals and organizations). The number of points within the system that must be addressed in order to affect the system defines its complexity. Basic beliefs hold that there are no complex systems in reality and that there are no contradictions in reality. It is assumed that underlying any contradiction, there must be at least one erroneous assumption.
Tools for thinking
The Theory of Constraints Thinking Processes are tools that enable use of cause-and-effect logic and necessity to first, gain an understanding of reality and why it is the way it is, and to then find ways of improving it by altering the currently prevailing assumptions and causalities. The two basic constructs are: causality (If… then…), and necessity (In order to… I must…).
So with this theory and GM Research’s software, Kohls’ team implemented the plan and saw throughput increase significantly, while overtime decreased. It took some time, since many things typically must be fixed before improvements are seen. Results were so good, however, that the organization looked for a way to implement TOC practices across many plants.
Working in the Detroit/Hamtramck plant, Kohls’ team members had been able to achieve buy-in from all of the affected departments. Going into other plants with the idea proved to be harder. Kohls went to the central office group and pitched the idea of how to improve throughput based on the experiences at Hamtramck. Immediately, people from every other plant objected, saying that their plants were all so much different from that one that the same methodology wouldn’t work. The question became, how to find a way to show each plant that the ideas were indeed applicable to them.
The answer once again came from the Goldratt Institute in a program called Jonah training. It outlines the thinking processes and overall framework of the TOC in a manner that allows participants to prove for themselves that the method really works. According to the theory, problems are really a result of not understanding the process. People in each of the plants have to internalize the tool. Training uses methods such as game playing and computer simulation to encourage students to learn for themselves how bottlenecks occur, and then to learn techniques to solve them.
The off-site training was valuable, says Kohls, because trying to do all this on the plant floor would have placed too much pressure on the participants and potentially interfered with production. By approaching the project with the understanding of why people don’t accept new ideas, confidence grew and spread throughout the company. Over time, as the process improved and successes were documented, GM approved putting a “throughput engineer” in each plant.
“Letting people experience the thinking process is how we got them to buy in,” says Kohls.
He adds, “We found that most problems are caused by core conflicts like finger pointing and not working as teams. Realistically, we just didn’t have enough discipline. We came to the realization that making money has to be one of the things we all are concerned about. But that is hard to put into practice. What people really think about is the minimum amount of work they need to do to get people off their backs. Instead, people need to make a financial decision. I could say that I improved efficiency of the machine, but then wonder why that didn’t improve the bottom line. As simple as that sounds, it is the essential point. Supervisors and managers enforce what we do, and they need to understand that financial criteria are the most important.”
Simulation helpful
A key part of analysis and training provided by the Goldratt Institute was the use of simulation. AutoMod from Salt Lake City-based Brooks-PRI Automation was chosen for this crucial step. AutoMod is a personal computer (PC)-based application that combines algorithms, math and graphics to represent reality. Simulation applications build a model of the real system, and then exercise the model to understand the behavior of the system and evaluate different strategies. In general, simulation is a good tool for evaluating capacity, testing design ideas and obtaining process statistics.
It is especially useful for processes in which the product is seen, such as assembly lines, warehousing and machining lines. Sometimes a simulation program can incorporate control programs and emulate actual automation. In these cases, simulation would be used to model throughput and allow both students and engineers to try different scenarios of bottleneck reduction in order to predict impact on total throughput.
Why predict throughput? Net profit equals selling price minus raw material cost times throughput minus operation expenses. The elements of this equation are generally well understood, with the exception of throughput. Because one objective of Kohls’ process is to use financial data as an aid to engineering decision making, predicting throughput is essential in order to predict plant profitability.
Because analysis in the TOC model emphasizes knowing current reality with hard data, it seemed to Kohls’ group that it should be easy to just plug the numbers into the simulation program and get the answer. But there were problems with this step, as well. One thing Kohls’ group discovered was that asking the same thing of two different simulators could result in different answers, due to use of different tools and processes. It was better to build on and then refine just one tool to maintain consistency.
Using simulation successfully requires discipline to use data that reflects actual plant performance, then ensuring that new systems perform well in simulation before implementing.
Combine objective analysis of plant production statistics, TOC Thinking Process, hard data and simulation of reality to reach the goal. And just what is the goal? To make money, now and in the future.
Keys to success
As Kevin Kohls and his group implemented Theory of Constraints Throughput Analysis at General Motors, they discovered five keys essential to the success of this type of project:
• Make the goal clear.
• Emphasize global measures of net profit and return on net assets (RONA).
• Emphasize throughput, since it has huge impact on profits.
• Recognize the importance of accurate data and simulation to test theories.
• Focus analysis tools on throughput.
Six Steps to Achieving Buy-In
Goldratt Institute analysts have determined that there are six steps that must be climbed to achieve success in implementing a Theory of Constraints project. These are:
1. Secure agreement on the problem to be solved.
2. Secure agreement on the direction of a solution.
3. Verify that the proposed solution will deliver the desired results.
4. Ensure that all significant potential negative side effects have been identified and prevented from happening.
5. Identify and address all significant potential obstacles that could block implementation of the solution.
6. Ensure that all the necessary leadership is committed to making the implementation successful.
Source: Goldratt Institute
Leaders relevant to this article: