hellooo i need a summary, thanks.

Controls Follow Strategy

Whatever

is not essential to the attainment of a companys objectives should bemeasured infrequently and only to prevent deterioration. It should be strictly con-trolled by exception. A standard should be set, measurement should be periodi-cal and on a sample basis, and only significant shortfalls below the establishedstandard should be reported.That we can quantify something is no reason for measuring it. The question is,Is this what a manager should consider important? Is this what a managers at-tention should be focused on? Is this the proper focus for controlthat is, for ef-fective direction with maximum economy of effort?3.

Controls have to be appropriate to the character and nature of the phenomenon mea- sured.

This may well be the most important specification; yet, it is least observedin the actual design of controls.Because controls have such an impact, it is important that we select not onlythe right ones but also the appropriate ones, to enable controls to give right vision

326 MANAGERIAL SKILLS

and to become the ground for effective action. The measurement must present theevents measured in

structurally true form

. Formal validity is not enough.Formal complaints or grievances coming out of a workforce are commonly re-ported in this form, five grievances per thousand employees per month. This isformally valid. But is it structurally valid? Or is it misdirection? The impressionthis report conveys is, first, that grievances are distributed throughout the work-force in a random manner. And second, the report gives the impression that theyare a minor problem, especially if we deal with five grievances per thousand em-ployees per month. This, while formally valid, completely

misrepresents and misin- forms

.Grievances are a social event. And social events are almost never distributed inthe normal distribution we find in the physical world. In this case, the greatmajority of departments in the plant, employing 95 percent of the workforce, nor-mally do not have even a single grievance during one year. But in one department,employing only a handful of people, we have a heavy incidence of grievancessothe five per thousand may well mean (and in the actual example from which Itook these figures, did mean) a major grievance per person per month. If this de-partment happens to be the final assembly through which all the production hasto pass, and if the workers in this department go out on strike when their griev-ances are being neglected by a management that has been misled by its own con-trols, the impact can be shattering. In this case, it bankrupted the company.Most measurements of sales performance, whether of the entire sales force or ofthe individual salesperson, report sales in total dollars. But in many businesses thisis an inappropriate figure. The same dollar volume of sales may mean a substantialprofit, no profit at all, or a sizable lossdependent on the product mix sold. Anabsolute sales figure not related to product mix, therefore, gives no control what-everneither to the individual salesperson, nor to the sales manager, nor to topmanagement. These are elementary things. Yet few managers seem to know them.The traditional information systems (especially traditional accounting) conceal ap-propriateness rather than highlight it. Without controls that bring out clearlywhat the

real structure of events is,

the manager lacks knowledge and therefore willtend to do the wrong things. For all the weight of the daily work pushes him orher toward allocating energies and resources in proportion to the

number

of events.There is a constant drift toward putting energies and resources where they canhave the least results, that is, on the vast number of phenomena which, together,account for practically no effects.4.

Measurements have to be congruent with events measured.

Alfred North White-head (18611947), the distinguished logician and philosopher, used to warn againstthe danger of the false concreteness. A measurement does not become more ac-curate by being worked out to the sixth decimal when the phenomenon is only

327Controls, Control, and Management

capable of being verified within a range of 50 to 70 percent. This is false concrete-ness, and misleading.It is an important piece of information that this or that phenomenon cannot bemeasured with precision but can be described only within a range or as a magni-tude. To say, We have 26 percent of the market, sounds reassuringly precise. Butit is usually so inaccurate a statement as to be virtually meaningless. What it re-ally means, as a rule, is We are not the dominant factor in the market, but we arenot marginal either. And even then the statement is no more reliable than thedefinition of the market that underlies it.It is up to the manager to think through what kind of measurement is appro-priate to the phenomenon it is meant to measure. He has to know when

approxi-mate

is more accurate than a firm-looking figure worked out in great detail. Hehas to know when a range is more accurate than even an approximate single figure.He has to know that

larger

and

smaller, earlier

and

later, up

and

down,

are quantita-tive terms and often more accurate, indeed more rigorous, than any specific figuresor range of figures.5.

Controls have to be timely

. Frequent measurements and very rapid reportingback do not necessarily give better control. Indeed, they may frustrate control.The time dimension of controls has to correspond to the time span of the eventmeasured.It has lately become fashionable to talk of real-time controls, that is, of con-trols that inform instantaneously and continuously. There are events where real-time controls are highly desirable. If a batch of antibiotics in the fermentationtank spoils as soon as temperature or pressure deviate from a very narrow range formore than a moment or two, real-time monitoring on a continuous basis is obvi-ously needed. But few events need such controls. And most cannot be controlledby them. Real time is the wrong time span for real control.Children planting a garden are so impatient, it is said, that they tend to pullout the radishes as soon as their leaves show, to see whether the root is forming.This is real-time controlmisapplied.Similarly, the attempt to measure research progress all the time is likely toconfound research results. The proper time span for research is a fairly long one.Once every two or three years, research progress and results should be rigorouslyappraised. In between such appraisals, an experienced manager keeps in touch. Heor she watches for any indication of major unexpected trouble, and, even more, forany sign of unexpected breakthroughs. But to monitor research in real timeassome research labs have been trying to dois pulling up the radishes.There is also the opposite danger, of not measuring often enough. It is particu-larly great with developments that (a) take a fairly long time to have results, and(b) have to come together at a point in the future to produce the desired end result.

328 MANAGERIAL SKILLS

6.

Controls need to be simple.

Every major New York commercial bank worked inthe 1960s on developing managerial controls, especially of costs and of allocationof efforts. Everyone spent a great deal of time and money on the task and came upwith control manuals. In only one of the banks were the manuals being used.When the executive in that bank was asked how he explained this, he did not (ashis interviewer expected him to) credit a massive training program or talk abouthis philosophy. He said instead, I have two teenage daughters. They know noth-ing about banking and are not terribly good at figures. But they are bright.Whenever I had worked out an approach to controlling an activity, I took my in-tended procedure home in draft form and asked my girls to let me explain it tothem. And only when I had it so simple that they could explain back to me whatthe procedure was intended to accomplish and how, did I go ahead. Only then wasit simple enough.Complicated controls do not work. They confuse. They misdirect attentionaway from what is to be controlled, and toward the mechanics and methodology ofthe control. If the user has to know how the control works before he can apply it,he has no control at all. And if he has to sit down and figure out what a measure-ment means, he has no control either.7.

Finally, controls must be operational.

They must be focused on action. Actionrather than information is their purpose. The action may be only study and analy-sis. In other words, a measurement may say, What goes on we dont understand;but something goes on that needs to be understood. But it should never just say,Here is something you might find interesting.This then means that controlswhether reports, studies, or figuresmust al-ways reach the person who is capable of taking controlling action. Whether theyshould reach anyone elseand especially someone higher upis debatable. Buttheir prime addressee is the manager or professional who can take action by virtueof his or her position in the flow of work and in the decision structure. And thisfurther means that the measurement must be in a form that is suitable for the re-cipients needs.Workers and first-line supervisors should receive measurements and controlinformation that enables them to direct their own immediate efforts toward re-sults they can control. Instead, typically, the first-line supervisor receives eachmonth a statement of the quality control results for the entire plant and theworker receives nothing. And top management usually receives the informationand measurements operating middle managers need and can use, and little ornothing of pertinence to their own top-management job.The reason for this is largely the confusion between control as domination ofothers and control as rational behavior.

Unless controls are means toward the latter, andthis means toward self-control, they lead to wrong action

. They are miscontrol.