The Goal: A Process of Ongoing Improvement, Third Revised Edition

by Eliyahu M. Goldratt

  until further notice they should not promise any delivery in less

  than four weeks from receipt of the order. It will jeopardize their

  new campaign, but that's life."

  Right in front of our eyes the baton has been passed. It's

  obvious who is the boss now. I feel proud and jealous at the same

  time.

  "Bob has taken over very nicely," Lou says as we enter my

  office. At least this front is covered."

  "Yes," I agree. "But I hate to put him in a position where his first independent actions are so negative."

  "Negative?" Lou asks. "What do you mean by negative?"

  "All the actions he is forced to take are leading in the wrong

  direction." I answer. "Of course, he doesn't have any choice, the alternative is much worse, but still. . . ."

  "Alex, I'm probably thicker than usual today, but I really

  don't understand. What do you mean by 'leading in the wrong

  direction?' '

  "Don't you see?" I'm irritated by the whole situation. "What is the unavoidable result of telling sales that they should quote

  four weeks' delivery? Remember, just two weeks ago we went out

  of our way to persuade them to quote two weeks. They didn't

  have much confidence then. Now, it will cause them to drop the

  entire sales campaign."

  "What else can we do?"

  "Probably nothing. But this doesn't change the end result;

  future throughput is down."

  "I see," says Lou. "On top of it, overtime is up significantly; putting the plant to work on the weekend will consume the entire

  overtime budget for the quarter."

  "Forget the budget," I say. "When Bob has to report it, I'll be the divisional president. The increased overtime is increasing operating expense. The point is that throughput will be down, op-

  erating expense will be up and increasing the buffers means that

  inventory will be up. Everything is moving in the opposite direc-

  tion of what it should."

  E.M. Goldratt

  The Goal: A Process of Ongoing Improvement

  Captured by Plamen T.

  333

  "Yup," he agrees.

  "Somewhere, I've made a mistake," I say. "A mistake that

  now is causing us to pull back. You know Lou, we still don't know

  what we're doing. Our ability to see what's in front of us resem-

  bles that of moles. We're reacting rather than planning."

  "But you've got to agree that we are reacting much better

  than before."

  "That's not a real comfort Lou, we're also moving much

  faster than before. I feel as if I'm driving looking only in the rear

  view mirror, and then, when it's almost too late, we make last

  minute course corrections. It's not good enough. It is definitely

  not good enough."

  E.M. Goldratt

  The Goal: A Process of Ongoing Improvement

  Captured by Plamen T.

  334

  40

  I'm driving back from headquarters with Lou. We've been

  doing this every day for the last two weeks. We are not in what

  one might call a cheerful mood. Now we know every little detail

  of what's going on in the division, and the picture doesn't look

  good at all. The only bright spot is my plant. No, I should get

  used to the fact that it's Donovan's plant. And it's not a bright

  spot, that's a gross understatement. It's the real savior.

  Donovan succeeded getting everything under control before

  the clients had any reason to complain. It will take him some time

  to regain the confidence of our sales people, but with me pressing

  from the other side it will not take long before it will be okay.

  This plant is so good that Lou and I were led astray for some

  time. The reports on the division gave us the impression that the

  situation is quite good. Only when we went through the elaborate

  work of separating out Donovan's plant was the real picture ex-

  posed. And it's not pretty. It's actually quite disastrous.

  "Lou, I think we did the exact thing that we knew we

  shouldn't do."

  "What are you talking about?" he says. "We haven't done

  anything yet."

  "We have gathered data, tons of data."

  "Yes, and there's a problem with the data," he says. "Frankly, I've never seen such a sloppy place. Every report is missing at

  least back-up details. You know what I found today? They don't

  even have a report on late receivables. The information is there

  but—can you believe—it's scattered in at least three different

  places. How can they operate this way?"

  "Lou, you're missing the point."

  "Am I? Do you know that with proper attention we can re-

  duce the open receivables by at least four days?"

  "And that will save the division," I say sarcastically.

  "No," he grins. "But it will help."

  "Will it?"

  When Lou doesn't answer I continue, "Do you really believe

  it will help? Look Lou, what have we learned? What did you

  yourself say when you asked for the job? Do you still remember?"

  E.M. Goldratt

  The Goal: A Process of Ongoing Improvement

  Captured by Plamen T.

  335

  Irritated he says, "I don't know what you're talking about.

  Don't you want me to correct things which are obviously wrong?"

  How am I going to explain it to him? I try again.

  "Lou, suppose that you do succeed in collecting four days

  out of the open receivables. By how much will throughput, inven-

  tory, and operating expense be improved?"

  "They'll all be slightly improved," he says. "But the major impact will be on cash. You shouldn't sneeze at four days' cash.

  Besides, improving the division requires many small steps. If ev-

  eryone does his share, together we can lift it."

  I drive silently. What Lou said makes sense, but somehow I

  know that he is wrong. Deadly wrong.

  "Lou, help me here. I know that improving the division will

  require many small improvements, but . . ."

  "But what?" he says. "Alex, you are too impatient. You know what they say, Rome was not built in a day."

  "We don't have hundreds of years."

  Lou is right, I am impatient. But shouldn't I be? Did we save

  our plant by being patient? And then I see it. Yes, many small

  actions are needed, but that doesn't mean that we can afford to

  be satisfied with actions that improve the situation. We must care-

  fully choose which ones to concentrate on, otherwise. . . .

  "Lou, let me ask you. How much time will it take you to

  change, for internal purposes only, the way that we evaluate in-

  ventory?"

  "The mechanical work is not a real problem, that won't take

  more than a few days. But if you're referring to the work it'll take

  to explain the ramifications, to explain to managers how this af-

  fects their day-to-day decisions, that's a different story. With con-

  centrated effort, I'd say it'll take weeks."

  Now I'm on solid ground.

  "What, do you think, is the impact of the way we currently

  evaluate inventory on the levels of finished stocks that the divi-

  sion currently holds."

  "Significant," he says.

  "How significant," I press. "Can you give me a number?"

  "I'm afraid not. Not even a meaningful evaluation
."

  "Let's try to do it together," I say. "Have you noticed the increase in finished goods that the division is holding?"

  "Yes, I have," he answers. "But why are you surprised? It's exactly what should be expected. Sales are down and the presE.M. Goldratt

  The Goal: A Process of Ongoing Improvement

  Captured by Plamen T.

  336

  sure to show profits is up, so they build finished goods inventory

  to generate fictitious inventory profits. I see what you mean. We

  can take the increase in finished goods as an indicator of the

  impact of the way we value inventory. Wow, it's about seventy

  days!"

  "Lovely," I say. "Compare it to your four days of receivables.

  On what should you work? Moreover," I keep on hammering,

  "what is the impact on throughput?"

  "I don't see any," he answers. "I see very clearly the impact on cash, on inventory, and on operating expense, but not on

  throughput."

  "Don't you?" I say mercilessly. "What was the reason that

  they gave us for not introducing the new models? Can you re-

  call?"

  "Yes," he says slowly. "They are convinced that introducing the new models will force them to declare all the old ones they're

  holding in stock as obsolete. That would cause a major blow to

  the bottom line."

  "So, we continue to offer the old stuff rather than the new.

  We continue to lose market share, but it's better than to bite the bullet of write-offs. Do you understand now the impact it has on

  throughput?"

  "Yes, I do. You are right. But Alex, you know what? With

  some extra effort I think that I can handle them both. I can work

  on the problem of the way we value inventory and at the same

  time arrange for more attention to the receivables."

  He still doesn't get it but now I think I know how to handle

  it.

  "What about the plant indicators?" I ask him.

  "That's a real Pandora's box," he sighs.

  "What is the damage there? Slightly bigger than four days?

  And what about the fact that sales continue to judge opportuni-

  ties according to the formal 'product cost' and desirable margins.

  Or even worse, that they will look for anything they can sell above

  variable cost. What's the damage there? And what about the

  transfer prices between us and the other divisions; that's a real

  killer. Do you want more?"

  "Stop, stop," he raised his hands. "You made your point. I

  guess I was inclined to deal with the open receivables issue just

  because there I know what to do, while in all the others . . ."

  "Afraid?" I ask.

  E.M. Goldratt

  The Goal: A Process of Ongoing Improvement

  Captured by Plamen T.

  337

  "Frankly, yes."

  "So am I, so am I." I mutter. "Where do we start? Where do

  we continue? On what should we concentrate first, on what sec-

  ond? It's overwhelming."

  "We need a process," he says. "That's obvious. It's too bad that the five-step process that we developed turned out to be

  false. No . . . Wait a minute Alex, that's not the case. At the end,

  the problem was not wandering bottlenecks. It was insufficient

  protection for the existing bottlenecks. Maybe we can use that

  five-step process?"

  "I don't see how, but it's worthwhile to check it. Should we

  head to the plant and give it a try?"

  "Certainly. I'll have to make some phone calls, but it's no

  problem."

  "No," I say. "I have some commitments for tonight."

  "You're right," he says. "It's very important but not urgent.

  It can wait for tomorrow."

  "Identify the system's constraint(s)," Lou reads from the

  board. "Do we accept it as the first step?"

  "I don't know," I say. "Let's examine the logic that brought us to write it. Do you remember what it was?"

  "Roughly," he says. "It was something about the fact that we adopted throughput as the number-one measurement."

  "I'm afraid that roughly is not good enough," I say. "At least not at such an early stage in our analysis. Let's try again, from

  first principles."

  "I'm all for it," he groans, "But what do you call first principles?"

  "I don't know. Something basic that we accept without hesi-

  tation."

  "Fine. I have one for you. Every organization was built for a

  purpose. We haven't built any organization just for the sake of its

  mere existence."

  "Correct," I laugh. "Even though I know some people in

  some organizations who seem to forget it."

  "Washington, you mean?"

  "That too. I thought about our corporation, but who cares.

  Let's keep going. Another basic fact is that any organization is

  comprised of more than one person, otherwise it's not an organi-

  zation."

  E.M. Goldratt

  The Goal: A Process of Ongoing Improvement

  Captured by Plamen T.

  338

  "Correct," says Lou. "But I don't see the point in all this. I can give you many more correct statements about organizations

  in general."

  "Yes, you probably can, but look at the conclusion that we

  can derive already. If any organization was built for a purpose

  and any organization is composed of more than one person, then

  we must conclude that the purpose of the organization requires

  the synchronized efforts of more than one person."

  "That makes sense," he says. "Otherwise we wouldn't need

  to create an organization; the efforts of individuals would suffice.

  So?"

  "If we need synchronized efforts," I continue, "Then the

  contribution of any single person to the organization's purpose is

  strongly dependent upon the performance of others."

  "Yes, that's obvious." With a bitter smile he adds, "Obvious to everybody except for our measurement system."

  Even though I wholeheartedly agree, I ignore his last com-

  ment. "If synchronized efforts are required and the contribution

  of one link is strongly dependent on the performance of the

  other links, we cannot ignore the fact that organizations are not

  just a pile of different links, they should be regarded as chains."

  "Or at least a grid," he corrects me.

  "Yes, but you see, every grid can be viewed as composed of

  several independent chains. The more complex the organization

  —the more interdependencies between the various links—the

  smaller number of independent chains it's composed of."

  Lou doesn't want to spend too much time on that point. "If

  you say so. But that's not so important. The important thing is

  you've just proven that any organization should be viewed as a

  chain. I can take it from here. Since the strength of the chain is

  determined by the weakest link, then the first step to improve an

  organization must be to identify the weakest link."

  "Or links," I correct. "Remember, an organization may be

  comprised of several independent chains."

  "Yes," he agrees impatiently. "But as you said, the complexity of our organizations almost guarantees that there are not

  many of them. In any event, it is taken care of by the S in paren-

  thesis that we put at the end of the word 'constraint'. Fine, Alex,

  what do we do about the measur
ements?"

  "Measurements?," I say in surprise. "Where did they come

  from?"

  E.M. Goldratt

  The Goal: A Process of Ongoing Improvement

  Captured by Plamen T.

  339

  "Didn't we agree yesterday that the distorted measurements

  are the biggest constraint of the division?"

  Bob Donovan is right. Lou certainly has a fixation on mea-

  surements. "They are definitely a big problem," I say carefully.

  "But I'm not convinced that they are the constraint . "

  "You're not?" Lou is astonished.

  "No I'm not," I say firmly. "Do you think that the fact that most of our products are already outdated in comparison to what

  the competition is offering is not a major problem? Don't you

  realize that the attitude in engineering, claiming that the basic

  rule of nature is that a project never finishes on time, is an even

  bigger problem. And what about marketing, have you seen any

  marketing plan that has any chance of turning the situation

  around?"

  "No," he grins. "As a matter of fact everything that I've seen of long term planning should be more appropriately categorized

  under 'long term bullshitting.' '

  I'm on a roll. Today asking me about problems is like open-

  ing a dam. "Wait Lou, I haven't finished. What about the mental-

  ity that is so prevalent in headquarters, the mentality of covering

  your ass. Haven't you noticed that whenever we asked about

  something that doesn't go so well, everyone almost automatically

  started to blame everybody else?"

  "How could I not notice. Okay, Alex, I get your point. There

  are major problems all over. It seems that in our division there is

  a whole herd of constraints, not just a few."

  "I still claim that there are only few constraints. Our division

  is too complex to have more than a very few independent chains.

  Lou, don't you realize that everything we mentioned so far is

  closely connected? The lack of sensible long-term strategy, the

  measurement issues, the lag in product design, the long lead

  times in production, the general attitude of passing the ball, of

  apathy, are all connected. We must put our finger on the core

  problem, on the root that causes them all. That is what actually is

  meant by identify the constraint. It's not prioritizing the bad ef-

  fects, it's identifying what causes them all."

  "How are we going to do that? How are we going to identify