These changes, though of an emergency nature, coincided with a sweeping reorganization of General Motors, going to the roots of industrial philosophy. The language of corporation minutes is laconic, but the consequences can be far reaching, as they were in this instance. The first business of the new administration, taken up at the last meeting of the old Executive Committee on December 30, 1920, is recorded as follows:
The President submitted for the consideration of the Executive Committee a new Organization Chart of our Corporation together with an explanatory letter, and same was discussed at length. (Note 3-1.)
This was unanimously approved and ordered sent to the board of directors, who also approved it. It was made effective on January 3, 1921.
The plan thus adopted was with modifications the one I had drafted about a year earlier under the title "Organization Study" and submitted to Mr. Durant for his consideration. (Note 3-2.) Since this plan has become the foundation of management policy in the modern General Motors—an expression of the basic principles of "decentralization" that govern its organization—and is said thereby to have had some influence on large-scale industrial enterprise in the United States, I shall say something here about its origin and substance.
First as to its origin. It has been supposed by some students that General Motors took its decentralized type of organization from the du Pont Company, as a result of the relationship of the two companies. Both managements at that time were in fact independently concerned with problems of organization, and both eventually adopted principles of decentralization. But they proceeded from opposite poles. The du Pont Company then was evolving from a centralized type of organization, common in the early days of American industry, while General Motors was emerging from almost total decentralization. General Motors needed to find a principle of co-ordination without losing the advantages of decentralization. These different backgrounds of the General Motors Corporation and the du Pont Company, together with the differences in nature and marketing of the products of the two enterprises, made it impractical for the same model of organization to serve properly for both of them.
Du Pont executives had been working on their own reorganization problem for a couple of years; but it was not until nine months after General Motors adopted its plan of organization that the du Pont Company also adopted a decentralized scheme. The two plans did not share their particulars, but only the management philosophy of decentralization.
The two types of operating problems, one arising from too much centralization (du Pont) and the other from too much decentralization (General Motors), were soon to be met by many large American manufacturing enterprises. One reason, perhaps, why General Motors and du Pont met and answered their organization problems early was that in 1920 and 1921 their operating problems were larger and more complex than those of most contemporary American industrial enterprises. I believe it is also true that we recognized the problems and thought more in terms of organizational principles and philosophies than did most businessmen of that time. The principles of organization got more attention among us than they did then in the universities. If what follows seems academic, I assure you that we did not think it so.
I wrote the "Organization Study" for General Motors as a possible solution for the specific problems created by the expansion of the corporation after World War I. I cannot, of course, say for sure how much of my thought on management came from contacts with my associates. Ideas, I imagine, are seldom, if ever, wholly original, but so far as I am aware, this study came out of my experience in Hyatt, United Motors, and General Motors. I had not been much of a book reader, and if I had been, I understand that I would not have found much in that line in those days to help; and I had no military experience. In the course of my twenty years or so at Hyatt I learned to operate a single industrial unit, relatively small in size and with one basic product. This unit contained the elementary functions of a manufacturing business: engineering, production, sales, and finance. But I had only a small board of directors, no executive committee, and no organization problems of the General Motors type.
In United Motors I met for the first time the problems of operating a multiple-unit organization with different products made by separate divisions. All that held United Motors together in its beginning was the concept of automotive parts and accessories. We made horns, radiators, bearings, rims, and the like, and we sold them to both automobile producers and the public. Certain limited areas of possible co-ordination presented themselves; for example, the servicing of the numerous small products made by the different divisions. Separate service agencies for such small items were uneconomic. I therefore set up a single, nationwide organization called United Motors Service, Inc., on October 14, 1916, to represent the divisions, with stations in twenty-odd large cities and several hundred dealers at other points. The divisions naturally resisted tins move for a while, but I persuaded them of the need for it, and for the first time learned something about getting decentralized management to yield some of its functions for the common good. The service organization is still operating in General Motors and has grown along with the business as a whole. I considered setting up a common laboratory for research, and most likely would have done so if we had not entered General Motors. I did establish in United Motors a unity of business purpose through the principle of return on investment. By placing each division on its own profitmaking basis, I gave the general office a common measure of efficiency with which to judge the contribution of each division to the whole. In this connection I devised a system of standard accounting which Albert Bradley, long-time chief financial officer of General Motors, later was kind enough to say was pretty good for a layman.
In the great expansion in General Motors between 1918 and 1920, I had been struck by the disparity between substance and form: plenty of substance and little form. I became convinced that the corporation could not continue to grow and survive unless it was better organized, and it was apparent that no one was giving that subject the attention it needed.
An example, close to home for me: When the United Motors group was brought into the General Motors Corporation in late 1918, I found that if I followed the prevailing practice of inter-corporate relations I would no longer be able to determine the rate of return on investment for these accessory divisions individually or as a group. This would necessarily mean that I would lose some degree of managerial control over my area of operations. At that time, material within General Motors was passing from one operating division to another at cost, or at cost plus some predetermined percentage. My divisions in the United Motors Corporation had sold both to outside customers and to their allied divisions at the market price. I knew that I operated a profit-making group, and I wished to continue to be able to demonstrate this performance to the general management, rather than to have my operating results on inter-divisional business swallowed up in the extra bookkeeping profits of some other division. It was a case of keeping the information clear.
It was not, however, a matter of interest to me only with respect to my divisions, since as a member of the Executive Committee, I was a kind of general executive and so had begun to think from the corporate viewpoint. The important thing was that no one knew how much was being contributed—plus or minus—by each division to the common good of the corporation. And since, therefore, no one knew, or could prove, where the efficiencies and inefficiencies lay, there was no objective basis for the allocation of new investment. This was one of the difficulties with the expansion program of that time. It was natural for the divisions to compete for investment funds, but it was irrational for the general officers of the corporation not to know where to place the money to best advantage. In the absence of objectivity it was not surprising that there was a lack of real agreement among the general officers. Furthermore, some of them had no broad outlook, and used their membership on the Executive Committee mainly to advance the interests of their respective divisions.
I had taken up the question of inter-divis
ional relations with Mr. Durant before I entered General Motors and my views on it were well enough known for me to be appointed chairman of a committee "to formulate rules and regulations pertaining to inter-divisional business" on December 31, 1918. I completed the report by the following summer and presented it to the Executive Committee on December 6, 1919. I select here a few of its first principles, which, though they are an accepted part of management doctrine today, were not so well known then. I think they are still worth attention.
I stated the basic argument as follows:
The profit resulting from any business considered abstractly, is no real measure of the merits of that particular business. An operation making $100,000.00 per year may be a very profitable business justifying expansion and the use of all the additional capital that it can profitably employ. On the other hand, a business making $10,000,000 a year may be a very unprofitable one, not only not justifying further expansion but even justifying liquidation unless more profitable returns can be obtained. It is not, therefore, a matter of the amount of profit but of the relation of that profit to the real worth of invested capital within the business. Unless that principle is fully recognized in any plan that may be adopted, illogical and unsound results and statistics are unavoidable . . .
There seems to me still to be no question about that. It is as I see it the strategic aim of a business to earn a return on capital, and if in any particular case the return in the long run is not satisfactory, the deficiency should be corrected or the activity abandoned for a more favorable one.
For sales to outside customers, I recognized in the report that the market would determine the actual price, and if this yielded a desirable return, the business in question might justify expansion. For exclusively inter-divisional transactions I recommended that the starting point should be cost plus some predetermined rate of return, but only as a guide. To avoid the possibility of protecting a supplying division which might be a high-cost producer, I recommended a number of steps involving analysis of the operation and comparison with outside competitive production where possible. The point I wish to make here relates not to technique—which other people know better than I do—but to the general principle of rate of return as the measure of the worth of a business. That idea was fundamental in my thinking about management problems.
On the influence of rate of return on decentralization and the relation of the part to the whole, I made several points of which the following seem to me to be of interest.
As to its bearing on organization:
. . . [It] Increases the morale of the organization by placing each operation on its own foundation, making it feel that it is a part of the Corporation, assuming its own responsibility and contributing its share to the final result.
As to its bearing on financial control:
. . . [It] Develops statistics correctly reflecting the relation between the net return and the invested capital of each operating division—the true measure of efficiency—irrespective of the number of other divisions contributing thereto and the capital employed within such divisions.
As to its bearing on strategic investment:
. . . [It] Enables the Corporation to direct the placing of additional capital where it will result in the greatest benefit to the Corporation as a whole.
So far as I know, tins was the first written statement of the broad principles of financial control in General Motors.
I continued thereafter to occupy myself with the subject of organization.
In the late summer of 1919 I had gone abroad with a group of General Motors executives to study the overseas prospects of the corporation. In the group were Mr. Haskell, its chairman, Mr. Kettering, Mr. Mott, Mr. Chrysler, Albert Champion, and Alfred T. Brandt, who acted as secretary. On our way over on the S.S. France we held regular meetings on overseas matters and at other times met and discussed problems of organization. I can recall that we had such conversations but I cannot recall what was said. Mr. Haskell, it appears, regarded them as being of some importance at the time. At the end of a letter to Mr. Durant on October 10, 1919, shortly after we returned to the United States, he wrote the following:
We started working on organization matters the day we left New York and the entire committee participated in conferences which resulted in an agreement, and a report is being prepared . . . which we believe will be workable and help to lighten all of our burdens. However, these matters can best be dealt with personally rather than in a perhaps already too long communication.
I do not know what Mr. Haskell meant by saying that we agreed, other than perhaps on the need for better organization; my recollection is that there was more disagreement than agreement. Nor do I know of any report on organization that came out of these discussions.
It is a long time to go back in memory and find the precise times and places of things, especially when their importance was not realized at the time. I have instituted searches to verify or correct my memory on origins. I find, for example, that during the year 1919, as a member of the Executive Committee, I along with others performed a number of tasks concerned with organization, and in connection with them developed in a rudimentary way some of the ideas that I wrote up in my general "Organization Study."
One of these tasks was the study of inter-divisional business, discussed above. Another was the study of appropriations-request rules, which is discussed in a later chapter. In the midst of this welter of thought and attempted action, and a half year before the actual economic and management crisis began, I drafted the "Organization Study" and circulated it unofficially. It became a kind of "best seller" in the corporation all during 1920; I received numerous letters from executives requesting copies of it, so many, in fact, that I found it necessary to reproduce it in quantity. It had no competition; that is, no other tangible effort was made, so far as I know, to achieve a general solution of the organization problem.
In September 1920 I sent a copy of the study to Pierre S. du Pont, then chairman of the corporation. We had an exchange about it. I wrote:
My dear Mr. DuPont:—
Referring to our conversation the other day, I am enclosing herewith copy of Organization Study worked out about a year ago.
I have reviewed this in the light of development since then and a greater insight in the working of our organization, and I do not see as I would recommend any radical changes except in reference to the additional recommendations . . . which I would not make as I view the situation now.
If you have time to read this Study at all, please bear in mind that it was worked out on a basis of what I thought would be acceptable to all interests rather than what would form an ideal organisation. If it were to be worked out along the latter line, I should be in favor of ultimately appointing an executive in charge of the three groups listed on Page 6; the miscellaneous group other than the Export Corporation and the Acceptance Corporation, ultimately passing into one of the other three groups. This would reduce the number of executives reporting directly to the President, to five, thus giving the President time to study into the broader problems.
Mr. du Pont replied:
Dear Mr. Sloan:
I am glad that you followed up our conversation with a copy of your study of a year ago. At the first opportunity, I shall read this carefully and hope for another discussion on the subject with you.
At the end of November 1920, when Mr. Durant went out and Mr. du Pont became president, the new administration needed a scheme of organization immediately. Mr. Durant had been able to operate the corporation in his own way, as the saying goes, "by the seat of his pants." The new administration was made up of men with very different ideas about business administration. They desired a highly rational and objective mode of operation. The "Organization Study" served the purpose and, as I have related, it was officially adopted, with some revision, as basic corporation policy.
The study was primitive by comparison with present-day knowledge of management. And it was written from the point
of view of presenting something that I thought would be acceptable to Mr. Durant. So it was not without constraints. It began as follows:
The object of this study is to suggest an organization for the General Motors Corporation which will definitely place the line of authority throughout its extensive operations as well as to co-ordinate each branch of its service, at the same time destroying none of the effectiveness with which its work has heretofore been conducted.
The basis upon which this study has been made is founded upon two principles, which are stated as follows:—
1. The responsibility attached to the chief executive of each operation shall in no way be limited. Each such organization headed by its chief executive shall be complete in every necessary function and enable[d] to exercise its full initiative and logical development.
2. Certain central organization functions are absolutely essential to the logical development and proper control of the Corporation's activities.
This does not need much interpretation. It asks first for a line of authority, co-ordination, and the retention of the effectiveness of the then prevailing total decentralization. But looking back on the text of the two basic principles, after all these years, I am amused to see that the language is contradictory, and that its very contradiction is the crux of the matter. In point l, I maximize decentralization of divisional operations in the words "shall in no way be limited." In point 2, I proceed to limit the responsibility of divisional chief executives in the expression "proper control." The language of organization has always suffered some want of words to express the true facts and circumstances of human interaction. One usually asserts one aspect or another of it at different times, such as the absolute independence of the part, and again the need of co-ordination, and again the concept of the whole with a guiding center. Interaction, however, is the thing, and with some reservation about the language and details I still stand on the fundamentals of what I wrote in the study. Its basic principles are in touch with the central problem of management as I have known it to this day.
My Years With General Motors Page 7