THE PRINCIPLE OF SLACK ROPES

Introduction

In July of last year (1986) at Boston University, a few of us spent a morning discussing the evaluation of manufacturing performance. No one was trying to teach anything to anyone. We were all simply reporting on work we had done, and seeking feedback. Yet from this meeting came an idea which I think the management community might want to share. I call it “the principle of slack ropes.”

Present at the meeting were Professors Jinichiro “Jim” Nakane and Takeshi Nagai of Tokyo’s Waseda University, Professors Tom Vollmann, Fred Nanni and myself from Boston University, and Tom O’Brien and Larry Utzig from General Electric Company’s Corporate Consulting Services. All contributed to the intellectual ferment that morning, and I trust none will be offended if this belated report fails to credit some individual contributions.

Jim Nakane’s presentation was the catalyst which led to the principle of slack ropes. Nakane reported on efforts to improve the evaluation of manufacturing performance in Japan. He defined the Japanese measures of factory performance as quality, reliability, dependability, flexibility, and cost. He defined dependability as schedule performance, and flexibility as made up of:

1. Lead time, the time to deliver a new order.

2. The number of different products producible in a week.

3. Response time to a change in total demand.

In Japanese industry, reported Nakane, there is a tendency to try to measure plants in a uniform way, and to pursue national goals. Presently, the goal is to reduce lead time. “We are finding,” he said, “that as we gain ground … as production lead time is reduced … our quality improves, our work-in-process inventory decreases, and our total costs also decrease.”

As background information, Nakane briefly described Japan’s thirty-year experience in pursuit of quality, during which time every other aspect of manufacturing performance also improved, but stated that quality is no longer the primary focus of their attempts to improve manufacturing performance. (Prof. Nagai later explained that this was because, “Quality is no longer a competitive weapon in the Japanese market, because everybody has it. Hence,” he said, “manufacturers must distinguish themselves on some other variable, with lead time the present choice.

The present Japanese objective, according to Nakane, is to achieve lead times of ½ to 1/3 of their historical levels.

After Nakane spoke, the implications of his words began to sink in, and I went to the chalkboard and drew this diagram (Figure 1):

There was a consistency in the Japanese story that I hadn’t noticed before, and I wanted to know whether I was the only one who’d overlooked it. “Professor Nakane,” I said, “you pursued quality for thirty years, and saw dependability, flexibility, reliability, and cost all improve along with it. Now you shift to lead time, and in pursuing its reduction as the sole focus of your manufacturing improvement efforts, you are achieving improvements in quality, dependability and cost, as well.”

“Could there be a principle at work here? It is as though you can’t improve just one measure of factory performance, because they are all linked, as by slack ropes. Pick an area for improvement, such as lead time, and start pulling on it. It moves a little, but not far before something stops it … a tug from another area of performance. Perhaps the work-in-process inventory is so great that improving lead time on one product causes others to be later, yielding no net gain. So you turn the organization’s attention to reducing W.I.P. And when someone asks, ‘Why are we working on inventory reduction?’ you respond, “Because it is necessary if we are to reduce lead time, which is our sole manufacturing goal.” And the questioner understands, and cooperates, because the importance of reducing inventory has been made clear. And as inventory is reduced, lead times can be reduced (and costs come down), but only until the next rope becomes taut.”

“Perhaps,” I said, “the key to Japan’s success has been singleness of purpose. The Japanese only work on one performance measure at a time.”

“But we can’t measure just one thing,” commented a participant. “That was the trap we were in when we only focused on cost. Our efforts at improving measurements have been to broaden them to encompass quality, delivery, employee turnover, grievance rates and the like.”

That remark sparked a lively discussion of the difference between measuring something and making it an objective. Are the two necessarily the same? We can measure to track progress toward an objective, but also must measure to maintain constraints. Tom O’Brien said, “We have a saying at GE, ‘You get what you measure.’” Someone replied, “Then measure everything. The Japanese do, and no backsliding is tolerated on any performance measure … but improvement is only sought in one area at a time.

As a corollary, it was observed that cost reductions, per se, had never been the goal in Japan, but were usually the explicit goal of manufacturing performance improvement programs in the U.S. From this idea came, “Cost is a resultant vector.” Cost, it would seem, should not be the single purpose. No one could prove that … but focusing on cost reduction seemed not to have worked in the U.S. (If cost is a resultant vector, then I should redraw the diagram of the principle of slack ropes as a see-saw: lift quality, dependability, etc. on one side, and lower costs on the other. Please see Figure 2.)

AFTERTHOUGHTS

Much more was discussed at the July meeting which should be written up and aired in management circles. But it is the slack ropes idea that has kept nagging at me. I’ve mentioned it to executives, to colleagues and to students, and they have been unanimous in their reaction to the effect that I’m not the last to have perceived the role of singleness of purpose in the success of the Japanese manufacturing system. So I decided to prepare this belated report, and now offer a few afterthoughts.

Singleness of purpose is hardly a new motivational concept. Athletic coaches have long used it to motivate championship teams. It wins wars. It builds empires. Is it appropriate to manufacturing performance improvement? The Japanese example suggests that it is … provided that the purpose chosen for single-minded pursuit is one people can pursue with honor, and one that can be operated on at many levels.

Several authors have noted that manufacturing performance improvement in Japan seems to come about in a series of small steps, and through patient attention to detail. Perhaps the principle of slack ropes is the mechanism by which such organizational learning occurs.

I am reminded of Wickham Skinner’s concept of “Key Task Focus”. Skinner wrote in his 1974 Harvard Business Review article “The focused factory” that a factory is inherently limited by its technology and its management systems, and cannot be “all things to all people”. Hence, a firm should study its competitive strategy to determine the most important competitive demands made on the manufacturing function … and then incorporate those in a Key Task around which to design and manage manufacturing. Skinner’s words paralleled the principle of slack ropes when he wrote, “Not only does [Key Task] focus provide punch and power, but it also provides clear goals that can be readily understood and assimilated by members of an organization. It provides, too, a mechanism for reappraising what is needed for success, and for readjusting and shaking up old, tired manufacturing organizations with welcome change and a clear sense of direction.” Doesn’t that ring true with the Slack Ropes story? (“Why are we working on inventory?” “Because inventory is blocking lead time reduction, our key task.” Oh, that makes sense!”)

SLACK ROPES VS. MBO

Another of my nagging afterthoughts has been the contrast between the principle of slack ropes and Management by Objectives (MBO) as I have seen the latter implemented in the U.S. In MBO systems each manager sits down periodically with his or her direct reports and prepares a list of objectives. “Improve these seven areas to the agreed degree, and the organization will benefit,” is the typical implication. But if each manager has five or so direct reports, then his or her sub-organization is going to be pursuing thirty-five or more objectives. Will they all be congruent, and not interfere with each other? And who’s to say what is to be worked on when? And who’s to say the manager’s list and timing will mesh with those of his or her peers? The problem of coordinating improvement efforts under such a system is complex, and rarely attacked.

Contrast that approach with this one: Our organization is good, but must get better if it is to remain competitive. We will compete in the future by remaining strong in quality, reliability, and dependability, but we will increase market share by reducing lead time. That is our goal. All managers and workers will share this single purpose, and on its achievement we will measure our organizational success.

Such an approach is self-organizing. Initially only the order clerks and production schedulers need act … by establishing measures of order throughput time and promise performance, and then shortening promises a bit. But soon a rope will draw taut, and the organization will follow the taut rope to learn what is preventing further shortening of lead time, and the problem area thus located will be worked on next. Perhaps process quality is still not good enough … or machine set-ups take too long … or we wait too long for tooling from vendors. But whatever is the immediate problem, it alone is where we shall focus our attention. Then we’ll shorten lead times again, and wait for the next rope to draw taut.

The management approach implied by the principle of slack ropes seems almost too simple. No massive data needs. No analytical bureaucracy. Only a well-chosen, competitively essential goal … and a policy of attacking anything which retards progress, step by step. Can such an approach work in America?

OPTIMISM AND DOUBT

The principle can work here. In March of 1986 a semiconductor manufacturer in Massachusetts was troubled by overly long lead times, and decided to attack them. The operations manager, influenced by Eliyahu Goldratt’s ideas on improving work flow, led the attack by cutting lot sizes in his plant to at most one shift’s work at each stage of the process. For example, an order of 36,000 units was broken up into 30 lots of 1200. (Conventional wisdom in manufacturing would have seen such a move as insane.) As it happened, shipments exceeded incoming orders for the first time in months … though the order rate held steady … and three weeks’ worth of product waiting at final test became three days’ worth by month-end.

But along the way there were many taut ropes, and some interesting second-order effects. Inspection personnel objected to the increase in the number of lots to sample. But when told, “We must do this to reduce lead times,” they cooperated. Second order effect: same size samples from smaller lots yielded higher outgoing quality.

The paperwork accompanying each lot had always been filled out manually at the stockroom. With so many short lots, this pencil-pushing became a bottleneck. The stockroom supervisor located an unused personal computer in the building, “liberated it”, and the next day the labeling of lots was done with computer-printed stickers. Second-order effect: shop employees and customers both found the labels easier to read, and the error rate was reduced. And thus the firm followed rope after rope.

I believe that the slack ropes mechanism will work anywhere that a purpose is persistently pursued. Therein lies the basis of my optimism. But I have my doubts as to the willingness and ability of American corporations to adopt and to pursue strategically critical purposes with persistence. For thirty years the Japanese followed W. Edwards Deming in pursuit of quality through statistical process control. The auto industry in this country discovered Deming in 1980. For a few years thereafter, I’ve been told, “You could get approval of about anything in the name of quality.” But in the July 28, 1986 issue of AutoWeek, George Damon Levy wrote:

“This story should have a happy ending. But will it?

You don’t hear so much about W. Edwards Deming these days in Detroit. And you don’t see so much effort – or progress – on the quality front from some American car makers.”

That quotation is frightening if, as it suggests, five years is the limit of consistently pursued purpose in the U.S.

DOES WHICH GOAL MATTER?

As some early readers have pointed out, tagging the central idea of this report a principle is presumptuous. The managerial pattern it implies has been seen to work, but only where quality improvement and lead time reduction have been chosen as goals. It may well be that those two measures of manufacturing are special. Does it matter which goal a manger chooses to define as the central purpose? Might some other performance measures behave as does cost, following but not leading the chain of general improvements I’ve described? Clearly, those questions and others require study before the term principle can be properly applied.

General principle or not, the self-ordering process of improvement in a complex managerial system implied by the slack ropes idea has kept generating afterthoughts. My final afterthought is hopeful. I’ve seen breakthroughs over the past two decades in America’s understanding of the competitive role of manufacturing, of the meaning of quality, and of the internal logic of production flow. We now have the tools. Let’s manage them on purpose.

Taken from the O.M. Review, Spring, 1987:

“Duncan C. McDougall is Associate Professor of Operations Management at Boston University where his research interests include Performance Measurements in Manufacturing and Manufacturing Strategy in Diversified Firms. In 1966 Mr. McDougall began his career with Chevrolet Flint Pressed Metal as a Materials Follow-up Man and Shipping Foreman. After a Harvard MBA in 1970, he was research assistant to Wickham Skinner for a year, then joined Rockwell International’s Draper Division, where he served as Facilities Planner, Manager of Business Planning, Plant Manager, and Product Line Manager. McDougall holds a doctorate in POM from Harvard University, and has spent the last eleven years teaching, first at Plymouth State College in New Hampshire, and since 1980 at BU. In May of 1987 he was a warded the Broderick Prize for Distinguished Teaching by his colleagues at Boston University’s School of Management.”

2004 Update: Having returned to New Hampshire in 1988 as president of a small manufacturing firm, Dr. McDougall is now Professor of Business Administration at Plymouth State University.