Critical Chain: A Business Novel

by Eliyahu M. Goldratt

Besides, there is another tack. When push comes to shove, we will reduce the targeted specifications of the project."

"Everybody knows that projects don't finish on time or on budget, and if they do it means they had to compromise on content. Especially in systems programming or product design."

"Can we conclude," I ask the class, "that the problems common to all projects are," and turning to the board I write as I continue to talk, "the high probability of, 1. Budget overruns; 2. Time overruns; and many times, 3. Compromising the content."

"There is a pattern here. The lower the level of the person, the more the finger points internally, rather than externally. You'll find the same thing in my report."

"The higher the uncertainty the longer the tail of the distribution," I remind them. "This is the median of the distribution," I draw the line on the graph. "It means that there is only a fifty percent chance of finishing at or before this time."

"Why not the median?" I ask him. "Because Murphy does exist," he laughs. "It also exists for Brian." "Come on," he says. "Only a suicidal, inexperienced person would choose the median."

"It especially makes sense because in most environments there is little positive incentive, if any, to finish ahead of time, but there are plenty

of explanations required when we are late. Under such conditions, I agree with Mark that almost nobody will choose an estimate they have a fifty percent chance of blowing. What probability will you feel comfortable with?" "Min...

"When we compare the time indicated by the median to the time you indicated as a reasonable estimate, it doesn't look like the safety you added is in the range of twenty percent." "Closer to two hundred percent," he

admits.

"Look at the graph," I urge them. "Do you understand that the time estimate that gives us a fifty percent chance is much shorter than the time estimate that provides an eighty percent chance of completing a step before the estimated time? And don...

"So, two hundred percent safety and more is the norm, not the exception," ...

"Which means that for each and every step of the project we factor in a lot of safety. Are you starting to realize," I ask the class, "the extent to which we do insert safety into our projects?"

"We concluded that the uncertainty existing in every project is the underlying main cause for most problems. Now we see that people are not blind to it and they do add a lot of safety in their planning. Do you agree that

we must investigate, more deeply, this issue of safety?"

"Maybe time estimates are a self-fulfilling prophesy?"

"What do you mean?" Mark is puzzled. "Remember what we learned in production?" Ruth asks.

"Steve told us that his plant got too many complaints about late deliveries; they had lousy due-date performance. So they started to promise their clients three weeks delivery time instead of two. That gave him the ability to release the work a week earlier."

"They said the work would take two weeks, it took two weeks plus. They added more safety time, and said that

it would take three weeks, it took three weeks plus. A self-fulfilling prophecy," Ruth summarizes.

"In production, most of the time parts spend in the plant they are waiting in queues in front of machines, or waiting for another part in front of assembly. Most of the lead time is not actual production, i...

"And if Dr. Silver is right, and each step in a project contains so much safety? What then? Then in projects also most of...

The effect is clear. Companies are so immersed in the mentality of saving money that they forget that the whole intention of a project is not to save money but to make money.

What is the critical path?"

"Critical path is defined as the longest chain of dependent steps. Longest in time, of course." "I know," he impatiently says. Then more slowly, "The critical path is the path through the steps of building the building, making it functional and installing the machines in it. A total of one hundred fifty days." "The critical path," I remind the class, "determines the time it will take to finish the project. Any delay on the critical path will delay the completion of the project. That's why the project manager must focus on it."

"Gantt charts, unlike PERT diagrams, involve decisions," I highlight to the class, "the decision of the planner when to start each path. Brian chose the late start for picking the vendors while Ted has chosen the early start."

"It's an optimization problem," Brian is confident. "We have to weigh the savings from postponing an investment against the chance of damage resulting from finishing the project a little late." One thing I passionately hate is optimization problems.

There are so many articles about these cases, all with involved mathematical models, all so tough and time-consuming to read. And from my experience, all have little practical use. But what can I do; it is an optimization problem.

To my surprise, she starts by saying, "I don't think that it's just a financial consideration. It's much more a management issue." "Explain." I try not to appear puzzled. "In a project there are many more paths than in our simplified example; many more entry steps." "Of course." "If we start all the paths at their earliest start, don't you think that the project leader will have too much on her hands? From my experience," she adds, "if I start too many things, I'm bound to lose focus, and losing focus is one thing a project leader cannot afford."

"If the project leader loses focus, the project is bound to be very late. The financial penalty of delaying the income from the completed project almost always dwarfs everything else."

"If we start a path on its late start," I wonder aloud, "then that path doesn't have any time slack. Which means that any delay on that path will also cause a delay in the project."

"Concentrating on everything is synonymous with not concentrating at all," I agree with him. "So where do we stand? If the project leaders use early starts, they will lose focus. If they use late starts, focusing is not possible at all. We have to find the mechanism, the rules, that will enable a project leader to focus."

"Maybe we can approach it from another angle? A proper control mechanism should keep us focused."

"Everybody knows what a control mechanism is: it measures the progress of the project. The problem is," he turns to me, "that by the time the progress report indicates something is wrong, it's usually too late."

Progress is measured according to the amount of work, or investment, already done, relative to the amount still to do. In all my students' cases, including the cases where milestones and progress payment were used, this measurement did not differentiate between work done on the critical path and work done on other paths.

"Can anybody predict the impact of measuring progress in this way?" I ask the class. "We reward starting each path at the earliest possible

time," Brian is quick to notice. "This measurement encourages the project leader to start unfocused." "Moreover," Charlie notices, "it encourages the project leader to continue being unfocused." "How come?" "Because according to our measurement," he explains, "progress on one path compensates for a delay on another. So we encourage progressing fast on one path even though another path is delayed." "What's bad about that?" Mark asks. "If I have difficulties in one path, why shouldn't I move on the other paths where I can?" "At the end they all merge together," Charlie reminds him. "All the advance that you gain in the open paths will have to wait for the delayed path a...

"A shortsighted project manager," Charlie's still talking to him, "can ignore the paths that are slowed down by problems, and the measurement will still indicate that the projec...

For a while. A long while. Only when the work is complete on all the other open paths, and only the problematic path remains, only then will the fallacy start to be revealed. Mark, don't see this as personal criticism. I do exactly the same ...

"Now I understand why so many projects take so long to complete their last ten percent. It's because, in measuring progress, we overlooked the importance of the critical path. I found the enemy, it is me. I'm the one who prepares all the project progress reports!"

"You claim that the main thing is the ability of the project manager to focus. You further claim that both early start and late start jeopardize the ability to focus, even though to different degrees. And then, you conclude that ignoring it is ignoring the heart of the problem."

"As long as we are unable to recommend when to start each path, it does have limited practical use," I admit. "You don't get it, do you?" Jim shakes his head. Apparently I don't. "Look Rick," he starts patiently, "you cannot describe focusing of a person by an equation." "So?" "So, this problem can't be solved mathematically." I still don't get it. "But it can be solved logically," I insist. "We now know where to look. If we keep on thinking we might find a logical procedure. It will be immensely helpful for projects."

"But, Jim, what about things like JIT and TQM? According to what you say, they are not academic enough. Still, we teach them in every university." "They've passed the test of reality."

Don't you see that the criteria we use to judge what is acceptable research are the problem. They choke us. It's almost impossible to put any meaningful contribution through the system. How can we . . ."

"How did you conclude that?" He is genuinely interested. "Did you use my criteria?" "Yes, of course." And I quote from memory, " ‘Criteria number one: Measurements should induce the parts to do what is good for the system as a whole.' Well, the way we measure the progress of a project, the measurement does almost the opposite. " ‘Criteria number two: Measurements should direct managers to the point that needs their attention.' In projects the measurement steers the project leader away. It's ridiculous, it's damaging, and in my opinion, it often leads to the failure of major projects. But

structured common sense,

Rick, like everybody else, has heard more and more about TOC in the last ten years or so. What he has heard and read made a lot of sense, but it kept changing. At first it was related to production scheduling. Then it became a banner to attack "product cost" methods. Then

marketing. Lately, it seems TOC is more connected with methods to remove friction between people. If Johnny could provide some order for this mess, it might be worthwhile sitting for a hour. Not much more than an hour. "My impression is," Johnny says, as he turns on one of the overhead projectors, "that TOC is a blend of three different, yet related, breakthroughs."