The Production Solution

Goldratt's first career was as a developer of computer software for factory management. He built a very successful business, and his clients were quite satisfied with the

Figure 2.9 The Throughput World/Cost World Evaporating Cloud exposes the manager's dilemma.

software; it gave them much-more-detailed information about where things were in their factories. He noticed after a while, however, that they were not making any more money using his software. He thought about this and realized that he had to derive the basic principle from a focus on the goal of a for-profit company (i.e., to make money now and in the future). The goal corresponds to what Deming means by the aim of a system.

Goldratt's books, most notably his initial international best seller The Goal, demonstrate how he invented and used the TOC to develop the elegant Drum-Buffer-Rope method for controlling production. The Drum-Buffer-Rope method is elegant because it is much simpler than the earlier methods of production management that attempted to control the production system through detailed complexity. The Drum-Buffer-Rope system focuses on the dynamics of the production system.

The drum is the processing capability of the constraint. It determines the overall throughput of the production process. Recall that throughput is the difference between sales revenue and raw material costs. To exploit (make maximum use of) the constraint in terms of throughput, you have to release the correct work into the system at the proper time in order not to ever starve the constraint and also not to overload it. Overloading the constraint (i.e., producing more than it can process) creates excess in process inventory (i.e., piles of incomplete work in front of the constraint). The rope transmits information from the drum to the release of work in order never to starve the constraint and to limit the build up of inventory.

Buffers are deliberate placement of in-process inventory to account for statistical fluctuations in the process system. Machines break, go out of alignment, or sometimes need unplanned maintenance. People do not always show up on time and do not work to a constant rate. The buffers account for these fluctuations.

Figure 2.10 illustrates a production system. Compare it to Figure 2.4 and note that this represents the inner workings of the overall business system depicted by

Process steps

Process steps

statistical fluctuations and dependent events.

Deming. Production is a subsystem of the overall business system, just as the circulatory system is subsystem of the human body.

Although Goldratt uses as a background in The Goal a factory that produces hardware products, the general nature of figures 2.4 works for any kind of system. The output is anything an organization does that it sends outside. Output includes scientific-research results, services of any kind, meetings, travel arrangements, reports, legal aid, software products, or any other output of any profit or non-profit organization. The systems mentioned also include government. Nonprofit and government systems obviously have a different goal (aim) than for-profit business.

Figures 2.4 and 2.10 are static pictures of a production system. The system stays fixed. Inputs flow through the system, converting to outputs. The flow through the system is not uniform. Each step in the processes has some amount of variation, often referred to as statistical fluctuation. Since workstations downstream of other workstations need the parts from the upstream workstations, they depend on the upstream workstations. This combination of dependent events and statistical fluctuations is an important issue in managing the overall system, especially at the constraint.

A system designed with capacity for steps upstream of the constraint equal to the capacity of the constraint cannot produce at the capacity of the constraint. The reason is that upstream fluctuations add up, leading to periodic starving of the constraint. The constraint can never make up this lost production because it is the constraint of the system. Therefore, all upstream workstations must have excess capacity in an optimum system.

Likewise, all workstations downstream of the constraint must have capacity that exceeds the capacity of the constraint. Otherwise, they can never make up any downside fluctuations in their performance relative to the performance of the constraint. Most of the time, they operate at the capacity of the constraint (the drum for the system), but the excess capacity allows them to catch up when necessary. This means all nonbottleneck machines in a production facility should spend some of their time not working.

This reasoning extends to the conclusion that a system operating with each step at optimum efficiency cannot be an efficient system. Most people intuitively believe that operating each part of a system at maximum efficiency causes the system to operate at maximum efficiency. You can see that an optimum system has to feed the bottleneck at its capacity and process the downstream parts at the bottleneck's average processing rate. This means that, on average, every nonbottleneck process must operate at lower efficiency than the bottleneck in order to have reserve capacity to make up for fluctuations.

This understanding is a major reason that the TOC is able to make such immediate impact, once people understand it. Managers design and operate most current systems without a critical understanding of the TOC. They work to cut costs everywhere, including the capacity of the constraint. They work to improve efficiency everywhere, including workstations upstream of the constraint that may cause the constraint to work on things that do not translate into short-term throughput. Once they understand the theory, identify the constraint, and improve its throughput, the system throughput increases immediately.

The computer systems that Goldratt was selling before he invented the TOC, as well as all other factory control systems, failed to account for the impact of the system constraint combined with statistical fluctuations and workstation dependency. Since the actual fluctuations are statistical, they are unpredictable. You can only predict the general behavior over a period of time for many items that flow through the system. Therefore, the schedules produced by the computer systems were out of date and incorrect as soon as they were produced. No wonder the scheduled did not cause the system to make more money. No wonder adding more detail to project plans did not make projects more successful.

In Critical Chain, Goldratt extended the concept of Drum-Buffer-Rope to project planning and performance. It is not a direct extension because project work on activities moves through time, while in a production facility, the parts move through fixed workstations. The same constraint phenomena apply to projects. The combination of statistical fluctuations and dependent events exists in a project. Current computer planning and control methods do not consider these fluctuations. Therefore, many of the same phenomena take place in projects that took place in production before Drum-Buffer-Rope (i.e., late delivery, longer and longer delivery times, resources not available when needed, and so on). More detailed planning or more sophisticated computer programs cannot correct these problems because of the structure of the project reality. You do not reduce uncertainty by cutting up tasks (remember the Fifth Discipline law about elephants). More-detailed plans increase static complexity but do not help deal with dynamic variation due to uncertain estimates.

For a project, the critical chain is the constraint. It is the focus for management of the system. The buffers are time buffers instead of material buffers. (Actually, in production the physical material buffers relate to time also. A pile of a certain size provides a certain period of protection for the machine that works on the pile.) Buffer management for projects is similar to the production counterpart. Counterparts to the rope are

• Release of projects to the system based on the ability of the constraint resource to process the project tasks;

• Release of activities for work based on the input from buffer reporting;

• The decisions made in buffer management on when to change the process (buffer recovery).

Many people have found it difficult to apply TOC understanding to their work. They can see from The Goal how to apply it to a physical production system but cannot see it in their system, which may be a service business, research-and-development (R&D) firm, a nonprofit organization, or a government agency. A middle manager of a current client recently stated, "Work in [our business domain] is way too complex. CCPM is doomed to fail given the inherent complexity of our work." (I have other clients successfully applying CCPM in organizations 40 times as large, with projects 100 times as complex.) There is no basis for a distinction based on the type of business: the theory applies to any business system and, so far, to every type of project people have tried it on—an extremely wide range of project types. You should expect more dramatic improvements for projects with greater uncertainty, but all organizations have projects with a range of uncertainty. It's Not Luck [25] shows how TOC tools apply to marketing, personal career planning, and personal issues at home.

Experience demonstrates that even in production systems, the constraint usually turns out to be a policy, not the physical bottleneck. The Goal [24] demonstrates this relative to financial and sales policies.

Consider a service business that answers telephone calls from customers. A common measure for such services is the number of calls per hour handled by each person. The goal of the system does not relate to the number of calls but to some effect from answering the calls (e.g., satisfied customers or orders taken). Calls have statistical fluctuations in their length, and they arrive at random times. Let us suppose you are a customer and want to order many things. Should the operator keep you on the line and get marked down for taking fewer calls per hour? How long will you wait for an operator to answer before you call a competitor?

As the manager of this service, how do you decide when to get more operators? If you have excess operators (so that the longer calls and variations in when calls arrive can be handled), your efficiency goes down, even though the throughput for the company may go up far more than the added operating expense. What is the constraint to this system?

Consider another case representative of many internal functions in a company, the human resource function. What is your department goal, and how does it relate to the company goal? How do you measure output to ensure that you are contributing to the company goal? Do you know where the company constraint is and how human resources might influence it? Goldratt defines several necessary conditions for achieving the goal of a company. One of these is to "Satisfy and motivate employees now and in the future." This condition directly affects the throughput of the company. Human resources clearly affect this necessary condition. Human resources also impact operating expenses in several ways, including their own contribution (cost) and their effect on company salaries and benefits through salary and benefit policies and union agreements.

Was this article helpful?

0 0
Project Management Made Easy

Project Management Made Easy

What you need to know about… Project Management Made Easy! Project management consists of more than just a large building project and can encompass small projects as well. No matter what the size of your project, you need to have some sort of project management. How you manage your project has everything to do with its outcome.

Get My Free Ebook

Post a comment