strategies for diversification

« Here is a method for measuring the profit potentialof alternative product-market strategies, starting with aforecast of trends and contingencies and then work-ing toward company needs and long-run objectives.

Strategiesfor Diversification

By H. Igor Ansoff

The Red Queen said, "Now, here, it takes allthe running you can do to keep in the same place.If you want to get somewhere else, you must runat least twice as fast as that!" ^

So it is in the American economy. Just to re-tain its relative position, a business firm mustgo through continuous growth and change. Toimprove its position, it must grow and change atleast "twice as fast as that."

According to a recent survey of the ioo larg-est United States corporations from 1909 to1948, few companies that have stuck to theirtraditional products and methods have grown instature. The report concludes: "There is noreason to believe that those now at the top willstay there except as they keep abreast in therace of innovation and competition." ^

There are four basic growth alternatives opento a business. It can grow through increasedmarket penetration, through market develop-ment, through product development, or throughdiversification.

A company which accepts diversification asa part of its planned approach to growth under-takes the task of continually weighing and com-paring the advantages of these four alternatives,selecting first one combination and then another,depending on the particular circumstances inlong-range development planning.

While they are an integral part of the over-

^ Lewis J. Carroll, Through the Looking-Glass (NewYork, The Heritage Press, 1941), p. 41.

" A. D. H. Kaplan, Big Enterprise in a CompetitiveSystem (Washington, The Brookings Institution, 1954),p. 142.

aU growth pattern, diversification decisions pre-sent certain unique problems. Much more thaiiother growth alternatives, they require a breakwith past patterns and traditions of a companyand an entry onto new and uncharted paths.

Accordingly, one of the aims of this articleis to relate diversification to the over-all growthperspectives of management, establish reasonswhich may lead a company to prefer diversifica-tion to other growth alternatives, and trace a re-lationship between over-all growth objectives andspecial diversification objectives. This will pro-vide us with a partly qualitative, partly quanti-tative method for selecting diversificatiian strate-gies which are best suited to long-term growth ofa company. We can use qualitative criteria toreduce the total number of possible strategies tothe most promising few, and then apply a returnon investment measure to narrow the choice ofplans still further.

Product-Market Alternatives

The term "diversification" is usually associ-ated with a change in the characteristics of thecompany's product line and/or market, in con-trast to niarket penetration, market development,and product development, which represent othertypes of change in product-market structure.Since these terms are frequently used inter-changeably, we can avoid later confusion by de-fining each as a special kind of product-marketstrategy. To begin with the basic concepts:

e The product line of a manufacturing companyrefers both to (a) the physical characteristics ofthe individual products (for example, size, weight,

113

114 Harvard Business Review

materials, tolerances) and to (b) the performancecharacteristics of the products (for example, an air-plane's speed, range, altitude, payload).

C In thinking of the market for a product wecan borrow a concept commonly used by the mili-tary — the concept of a mission. A productmission is a description of the job which theproduct is intended to perform. For instance, oneof the missions of the Lockheed Aircraft Corpora-tion is commercial air transportation of passengers;another is provision of airborne early warning forthe Air Defense Command; a third is performanceof air-to-air combat.

For our purposes, the concept of a mission ismore useful in describing market alternatives thanwould be the concept of a "customer," since acustomer usually has many different missions, eachrequiring a different product. The Air DefenseCommand, for example, needs different kinds ofwarning systems. Also, the product mission con-cept helps management to set up the problems insuch a way that it can better evaluate the per-formance of competing products.

€ A product-market strategy, accordingly, is ajoint statement of a product line and the corre-sponding set of missions which the products aredesigned to fulfill. In shorthand form (see EXHIBITi), if we let n represent the product line and fithe corresponding set of missions, then the pairof IT and ̂ is a product-market strategy.

With these concepts in mind let us turn nowto the four different t5^es of product-marketstrategy shown in EXHIBIT I :

C Market penetration is an effort to increasecompany sales without departing from an originalproduct-market strategy. The company seeks toimprove business performance either by increasingthe volume of sales to its present customers or byfinding new customers for present products.

C Market development is a strategy in whichthe company attempts to adapt its present productline (generally with some modification in the prod-uct characteristics) to new missions. An airplane

EXHIBIT I. PRODUCT-MARKET STRATEGIES FORBUSINESS GROWTH ALTERNATIVES

V.MARK.ETS

LINE ^

Wo

TT,

IT.

MARKETfenetraUon

O z

2 2U J

MARKET

DIVE

DEVELO

RSIFIC

'MENT

NATION1

company which adapts and sells its passenger trans-port for the mission of cargo transportation is anexample of this strategy.

e A produet development strategy, on the otherhand, retains the present mission and developsproducts that have new and different characteris-tics such as will improve the performance of themission.

C Diversification is the final alternative. It callsfor a simultaneous departure from the presentproduct line and the present market structure.

Each of the above strategies describes a dis-tinct path which a business can take towardfuture growth. However, it must be emphasizedthat in most actual situations a business wouldfollow several of these paths at the same time.As a matter of fact, a simultaneous pursuit ofmarket penetration, market development, andproduct development is usually a sign of a pro-gressive, well-run business and may be essentialto survival in the face of economic competition.

The diversification strategy stands apart fromthe other three. WhUe the latter are usuallyfollowed with the same technical, financial, andmerchandising resources which are used for theoriginal product line, diversification generallyrequires new skills, new techniques, and newfacilities. As a result, it almost invariably leadsto physical and organizational changes in thestructure of the business which represent a dis-tinct break with past business experience.

Forecasting GrowthA study of business literature and of com-

pany histories reyeals many different reasons fordiversification. Companies diversify to compen-sate for technological obsolescence, to distributerisk, to utilize excess productive capacity, to re-invest earnings, to obtain top management, andso forth. In deciding whether to diversify, man-agement should carefully analyze its futuregrowth prospects. It should think of marketpenetration, market development, and productdevelopment as parts of its over-all product strat-egy and ask whether this strategy should bebroadened to include diversification.

Long-Term TrendsA standard method of analyzing future com-

pany growth prospects is to use long-range salesforecasts. Preparation of such forecasts involvessimultaneous consideration of a number of majorfactors:

• General economic trends.

• Political and international trends.• Trends peculiar to the industry. (For ex-

ample, forecasts prepared in the airplane in-dustry must take account of such possibilitiesas a changeover from manned aircraft tomissiles, changes irt the government "mobili-zation base" concept with all that would meanfor the aircraft industry, and rising expendi-tures required for research and development.)

• Estimates of the firm's competitive strengthrelative to other members of the industry.

• Estimates of improvements in the companyperformance which can be achieved throughmarket penetration, market development, andproduct development.

• Trends in manufacturing costs.

Such forecasts usually assume that companymanagement will be aggressive and that manage-ment policies will take full advantage of the op-portunities offered by the different trends. Theyare, in other words, estimates of the best possi-ble results the business can hope to achieve shortof diversification.

Different patterns of forecasted growth areshown in EXHIBIT II , with hypothetical growthcurves for the national economy (GNP) and thecompany's industry added for purposes of com-parison. One of the curves illustrates a salescurve which declines with time. This may bethe result of an expected contraction of demand,the obsolescence of manufacturing techniques,emergence of new products better suited to themission to which the company caters, or otherchanges. Another typical pattern, frequentlycaused by seasonal variations in demand, is oneof cyclic sales activity. Less apparent, but moreimportant, are slower cyclic changes, such astrends in construction or the peace-war variationin demand in the aircraft industry.

If the most optimistic sales estimates whichcan be attained short of diversification fall ineither of the preceding cases, diversification isstrongly indicated. However, a company maychoose to diversify even if its prospects do, onthe whole, appear favorable. This is illustratedby the "slow growth curve." As drawn in EX-HIBIT II, the curve indicates rising sales which,in fact, grow faster than the economy as a whole.Nevertheless, the particular company may be-long to one of the so-called "growth industries"which as a whole is surging ahead. Such a com-pany may diversify because it feels that its pro-

Strategies for Diversification 115EXHIBIT II. TREND FORECASTS

SALESVOLUME

, ^ ' INDUSTRY

SLOW GROWTH

CYCLICGROWTH

DECLINE

I9S5 I960 1965

spective growth rate is unsatisfactory in com-parison to the industry growth rate.

Making trend forecasts is far from a precisescience. The characteristics of the basic envi-ronmental trends, as well as the efEect of thesetrends on the industry, are always uncertain.Furthermore, the ability of a particular businessorganization to perform in the new environmentis very difficult to assess. Consequently, anyrealistic company forecast should include sev-eral different trend forecasts, each with an ex-plicitly or implicitly assigned probability. Asan alternative, the company's growth trend fore-cast may be represented by a widening spreadbetween two extremes, similar to that shownfor GNP in EXHIBIT II .

ContingenciesIn addition to trends, another class of events

may make diversification desirable. These arecertain environmental conditions which, if theyoccur, will have a great effect on sales; however,we cannot predict their occurrence with cer-tainty. To illustrate such "contingent" events,an aircraft company might foresee these possi-bilities that would upset its trend forecasts:

• A major technological "breakthrough" whosecharacteristics can be foreseen but whose tim-ing cannot at present be determined, such asthe discovery of a new manufacturing processfor high-strength, thermally resistant aircraftbodies.

• An economic recession which would lead toloss of orders for commercial aircraft andwould change the pattern of spending formilitary aircraft.

• A major economic depression.• A limited war which would sharply increase

the demand for air industry products.


• A sudden cessation of the cold war, a currentlypopular hope which has waxed and wanedwith changes in Soviet behavior.

The two types of sales forecast are illustratedin EXHIBIT HI for a hypothetical company. Salescurves Si and S2 represent a spread of trendforecasts; and S3 and S4, two contingent fore-casts for the same event. The difference be-tween the two types, both in starting time andeffect oh sales, lies in the degree of uncertaintyassociated with each.

In the case of trend forecasts we can trace acrude time history of sales based on events whichwe fuUy expect to happen. Any uncertaintyarises from not knowing exactly when they willtake place and how they will influence business.In the case of contingency forecasts, we canagain trace a crude time history, but our uncer-tainty is greater. We lack precise knowledgeof not only when the event will occur but alsowhether it will occur. In going from a trendto a contingency forecast, we advance, so tospeak, one notch up the scale of ignorance.

In considering the relative weight we shouldgive to contingent events in diversification plan-ning, we must consider not only the magnitudeof their effect on sales, but also the relative prob-ability of their occurrence. For example, if asevere economic depression were to occur, itseffect on many industries would be devastating.Many companies feel safe in neglecting it intheir planning, however, because they feel thatthe likelihood of a deep depression is very small,at least for the near future.

It is a common business practice to put pri-mary emphasis on trend forecasts; in fact, inmany cases businessmen devote their long-rangeplanning exclusively to these forecasts. Theyusually view a possible catastrophe as "some-thing one cannot plan for" or as a second-order

EXHIBIT III. A HYPOTHETICAL COMPANY FORE-

CAST — NO DIVERSIFICATION

SALESVOLUME

time

correction to be applied only after the trendshave been taken into account. The emphasis ison planning for growth, and planning for con-tingencies is viewed as an "insurance policy"against reversals.

People familiar with planning problems inthe military establishment will note here an in-teresting difference between military and busi-ness attitudes. While business planning em-phasizes trends, military planning emphasizescontingencies. To use a crude analogy, a busi-ness planner is concerned with planning forcontinuous, successful, day-after-day operationof a supermarket. If he is progressive, he alsobuys an insurance policy against fire, but hespends relatively little time in planning for fires.The military is more like the fire engine com-pany; the fire is the thing. Day-to-day opera-tions are of interest only insofar as they can beutilized to improve readiness and fire-fightingtechniques.

Unforeseeable Events

So far we have dealt with diversification fore-casts based on what may be called foreseeablemarket conditions — conditions which we caninterpret in terms of time-phased sales curves.Planners have a tendency to stop here, to disre-gard the fact that, in addition to the events forwhich we can draw time histories, there is arecognizable class of events to which we canassign a probability of occurrence but which wecannot otherwise describe in our present stateof knowledge. One must move another notchup the scale of ignorance in order to considerthese possibilities.

Many businessmen feel that the effort is notworthwhile. They argue that since no informa-tion is available about these unforeseeable cir-cumstances, one might as well devote the avail-able time and energy to planning for the fore-seeable circumstances, or that, in a very generalsense, planning for the foreseeable also preparesone for the unforeseeable contingencies.

In contrast, more experienced military andbusiness people have a very different attitude.Well aware of the importance and relative prob-ability of unforeseeable events, they ask why oneshould plan specific steps for the foreseeableevents while neglecting the really important pos-sibilities. They may substitute for such plan-ning practical maxims for conducting one's busi-ness — "be solvent," "be light on your feet," "beflexible." Unfortunately, it is not always clear


proved position in their own industry may beidentified as companies that are notable for dras-tic changes made in their product mix and meth-ods, generating or responding to new competition.

"There are two outstanding cases in which theindustry leader of 1909 had by 1948 risen inposition relative to its own industry group and alsoin rank among the 100 largest — one in chemicalsand the other in electrical equipment. These two(General Electric and DuPont) are hardly recog-nizable as the same companies they were in 1909except for retention of the name; for in each casethe product mix of 1948 is vastly different fromwhat it was in the earlier year, and the marketsin which the companies meet competition are in-comparably broader than those that accounted fortheir earlier place at the top of their industries.They exemplify the flux in the market positionsof the most successful industrial giants duringthe past four decades and a general growth ratherthan a consolidation of supremacy in a circum-scribed line." *

This suggests that the existence of specificundesirable trends is not the only reason for di-versification. A broader product line may becalled for even with optimistic forecasts for pres-ent products. An examination of the foresee-able alternatives should be accompanied by ananalysis of how well the over-all company prod-uct-market strategy covers the so-called growthareas of technology — areas of many potentialdiscoveries. If such analysis shows that, becauseof its product lines, a company's chances of tak-ing advantage of important discoveries are lim-ited, management should broaden its technolog-ical and economic base by entering a number ofso-called "growth industries." Even if the de-finable horizons look bright, a need for flexi-bility, in the widest sense of the word, may pro-vide potent reasons for diversification.

Diversification Objectives

If an analysis of trends and contingencies in-dicates that a company should diversify, whereshould it look for diversification opportunities?

Generally speaking, there are three types ofopportunities:

(1) Each product manufactured by a companyis made up of functional components, parts, andbasic materials which go into the final assembly.A manufacturing concern usually buys a largefraction of these from outside suppliers. One way

* Ibid., p. 142.

to diversify, commonly known as vertical diversifi-cation, is to branch out into production of com-ponents, parts, and materials. Perhaps the mostoutstanding example of vertical diversification isthe Ford empire in the days of Henry Ford, Sr.

At first glance, vertical diversification seems in-consistent with our definition of a diversificationstrategy. However, the respective missions whichcomponents, parts, and materials are designed toperform are distinct from the mission of the over-all product. Furthermore, the technology in fabri-cation and manufacture of these parts and materialsis likely to be very different from the technologyof manufacturing the final product. Thus, verticaldiversification does imply both catering to newmissions and introduction of new products.

(2) Another possible way to go is horizontal di-versification. This can be described as the intro-duction of new products which, while they do notcontribute to the present product line in any way,cater to missions which lie within the company'sknow-how and experience in technology, finance,and marketing.

(3) It is also possible, by lateral diversification,to move beyond the confines of the industry towhich a company belongs. This obviously opensa great many possibilities, from operating bananaboats to building atomic reactors. While verticaland horizontal diversification are restrictive, in thesense that they delimit the field of interest, lateraldiversification is "wide open." It is an announce-ment of the company's intent to range far afieldfrom its present market stmcture.

Choice of Direction

How does a company choose among these di-versification directions? In part the answer de-pends on the reasons which prompt diversifica-tion. For example, in the light of the trends de-scribed for the industry, an aircraft companymay make the following moves to meet long-range sales objectives through diversification:

1. A vertical move to contribute to the techno-logical progress of the present product line.

2. A horizontal move to improve the coverageof the military market.

3. A horizontal move to increase the percen-tage of commercial sales in the over-all salesprogram.

4. A lateral move to stabilize sales in case of arecession.

5. A lateral move to broaden the company'stechnological base.

Some of these diversification objectives applyto characteristics of the product, some to those

Strategies for Diversification 117

EXHIBIT IV. CHANGES IN LIST OF THE IOO LARGEST INIJUSTRIAL CORPORATIONS

(even to the people who preach it) what thisflexibility means.

An interesting study by The Brookings Insti-tution ^ provides an example of the importanceof the unforeseeable events to business. EX-HIBIT IV shows the changing make-up of thelist of the I oo largest corporations over the last50 years. Of the 100 largest on the 1909 list(represented by the heavy marble texture) only36 were among the 100 largest in 1948; justabout half of the new entries to the list in 1919(represented by white) were left in 1948; lessthan half of the new entries in 1929 (repre-sented by the zigzag design) were left in 1948;and so on. Clearly, a majority of the giants ofyesteryear have dropped behind in a relativelyshort span of time.

Many of the events that hurt these corpora-tions could not be specifically foreseen in 1909.If the companies which dropped from the orig-inal list had made forecasts of the foreseeablekind at that time — and some of them musthave — they would very likely have found thefuture growth prospects to be excellent. Sincethen, however, railroads, which loomed as the

' A. D. H. Kaplan, op. cit.

primary means of transportation, have givenway to the automobile and the airplane; the tex-tile industry, which appeared to have a built-indemand in an expanding world population, hasbeen challenged and dominated by synthetics;radio, radar, and television have created meansof communication unforeseeable in significanceand scope; and many other sweeping changeshave occurred.

Planning for the UnknownThe lessons of the past 50 years are fully ap-

plicable today. The pace of economic and tech-nological change is so rapid that it is virtuallycertain that major breakthroughs comparable tothose of the last 50 years, but not yet foreseeablein scope and character, will profoundly changethe structure of the national economy. All ofthis has important implications for diversifica-tion, as suggested by the Brookings study:

"The majority of the companies included amongthe IOO largest of our day have attained their posi-tions within the last two decades. They are com-panies that have started new industries or havetransformed old ones to create or meet consumerpreferences. The companies that have not onlygrown in absolute terms but have gained an im-

of the product missions. Each objective is de-signed to improve some aspect of the balancebetween the over-aU product-market strategy andthe expected environment. The specific objec-tives derived for any given case can be groupedinto three general categories: growth objectives,such as I, 2, and 3 above, which are designedto improve the balance under favorable trendconditions; stability objectives, such as 3 and 4,designed as protection against unfavorabletrends and foreseeable contingencies; and flexi-bility objectives, such as 5, to strengthen thecompany against unforeseeable contingencies.

A diversification direction which is highly de-sirable for one of the objectives is likely to beless desirable for others. For example:

C If a company is diversifying because its salestrend shows a declining volume of demand, itwould be unvdse to consider vertical diversification,since this would be at best a temporary device tostave off an eventual decline of business.

€ If a company's industry shows every sign ofhealthy growth, then vertical and, in particular,horizontal diversification would be a desirable de-vice for strengthening the position of the com-pany in a field in which its knowledge and ex-perience are concentrated.

e If the major concern is stability under a con-tingent forecast, chances are that both horizontaland vertical diversification could not provide a suf-ficient stabilizing influence and that lateral actionis called for.

e If management's concern is with the narrow-ness of the technological base in the face of whatwe have called unforeseeable contingencies, thenlateral diversification into new areas of technologywould be clearly indicated.

Measured Sales GoalsManagement can and should state the objec-

tives of growth and stability in quantitativeterms as long-range sales objectives. This is il-lustrated in EXHIBIT V. The solid lines describea hypothetical company's forecasted perform-ance without diversification under a generaltrend, represented by the sales curve marked Si,and in a contingency, represented by Sa. Thedashed lines show the improved performance asa result of diversification, with S3 representingthe curve for continuation of normal trends andS4 representing the curve for a major reverse.

Growth. Management's first aim in diversify-ing is to improve the growth pattern of the com-pany. The growth objective can be stated thus:


Under trend conditions the growth rate of salesafter diversification should exceed the growth rateof sales of the original product line by a minimumspecified margin. Or to illustrate in mathematicalshorthand, the objective for the company in EX-HIBIT V would be:

S3 — S i ^ jO

where the value of the margin p is specified foreach year after diversification.

EXHIBIT V. DIVERSIFICATION OBJECTIVES

SALESVOLUME

T t-Ume

Some companies (particularly in the growthindustries) fix an annual rate of growth whichthey wish to attain. Every year this rate ofgrowth is compared to the actual growth duringthe past year. A decision on diversification ac-tion for the coming year is then based upon thedisparity between the objective and the actualrate of growth.

Stability. The second effect desired of diver-sification is improvement in company stabilityunder contingent conditions. Not only shoulddiversification prevent sales from dropping aslow as they might have before diversification,but the percentage drop should also be lower.The second sales objective is thus a stability ob-jective. It can be stated as follows:

Under contingent conditions the percentage de-cline in sales which may occur without diversifica-tion should exceed the percentage drop in saleswith diversification by an adequate margin, oralgebraically:

Si —̂ S2 S3 —̂ S4

Using this equation, it is possible to relatethe sales volumes before and after diversifica-tion to a rough measure of the resulting stability.Let the ratio of the lowest sales during a slumpto the sales which would have occurred in thesame year under trend conditions be called thestability factor F. Thus, F = 0.3 would meanthat the company sales during a contingency


amount to 30% of what is expected under trendconditions. In EXHIBIT VI the stability factor ofthe company before diversification is the valueFl = S2/S1 and the stability factor after diver-sification is F3 = S4/SS, both computed at thepoint on the curve where S2 is minimum.

Now let us suppose that management is con-sidering the purchase of a subsidiary. Howlarge does the subsidiary have to be if the parentis to improve the stability of the corporation asa whole by a certain amount? EXHIBIT VI showshow the question can be answered:

On the horizontal axis we plot the differentpossible sales volumes of a smaller firm that mightbe secured as a proportion of the parent's volume.Obviously, the greater this proportion, the greaterthe impact of the purchase on the parent's stability.

On the vertical axis we plot different ratios ofthe parent's stabihty before and after diversifica-tion (F3/F1).

The assumed stability factor of the parent is0.3. Let us say that four prospective subsidiarieshave stability factors of i.o, 0.9, 0.75, and 0.6.If they were not considerably higher than 0.3, ofcourse, there would be no point in acquiring them(at least for our purposes here).

EXHIBIT VI. IMPROVEMENT IN STABILITY FACTORAS A RESULT OF DIVERSIFICATION FOR FJ = 0.3

RATIC

. , F

18

11

F

) OF STABILITY fACTORSJ - STAB/J-/TY OF £>/i/£/?S/F/C4T/OA/

-ST-ABIL/TY B£FOR.B D/VBR.S/F/CAT/C

1

OT.

*»•

.•••«•

^ ^

^ ^

• ! : •

^ ^

^ *

a e O.8

==-• M B

/.o

F, -1.0

f, =0.9

a=0.75

Ft-as

SALES OF SUBSIDIARY AS A riSACTION OFPAR€NT SAL£S BEFORE DIVERSIFICATION

On the graph we correlate these four stabilityfactors of the subsidiary with (i) the ratio F3/F1and (2) different sales volumes of the subsidiary.We find, for example, that if the parent is todouble its stability (point 2.0 on the vertical axis),it must obtain a subsidiary with a stability of i.oand 75% as much sales volume as the parent, ora subsidiary vdth a stability of 0.9 and 95% of thesales volume. If the parent seeks an improvementin stability of, say, only 40%, it could buy acompany with a stability of 0.9 and 25% as muchsales volume as it has.

This particular way of expressing sales ob-jectives has two important advantages: (i) By

setting minimum, rather than maximum, limitson growth, it leaves room for the company totake advantage of unusual growth opportunitiesin order to exceed these goals, and thus providesdefinite goals without inhibiting initiative andincentive. (2) It takes account of the time-phasing of diversification moves; and since thesemoves invariably require a transition period, thenumerical values of growth objectives can beallowed to vary from year to year so as to allowfor a gradual development of operations.

Long-Range Objectives

Diversification objectives specify directionsin which a company's product-market shouldchange. Usually there will be several objectivesindicating different and sometimes conflictingdirections. If a company attempts to follow allof them simultaneously, it is in danger of spread-ing itself too thin and of becoming a conglom-eration of incompatible, although perhaps indi-vidually profitable, enterprises.

There are cases of diversification which havefollowed this path. In a majority of cases, how-ever, there are valid reasons why a companyshould seek to preserve certain basic unifyingcharacteristics as it goes through a process ofgrowth and change. Consequently, diversifica-tion objectives should be supplemented by astatement of long-range product-market objec-tives. For instance:

e One consistent course of action is to adopt aproduct-market policy which will preserve a kindof technological coherence among the differentmanufactures with the focus on the products ofthe parent company. For instance, a companythat is mainly distinguished for a type of engi-neering and production excellence would continueto select product-market entries which wouldstrengthen and maintain this excellence. Perhapsthe best known example of such policy is exempli-fied by the DuPont slogan, "Better things forbetter living through chemistry."

C Another approach is to set long-term growthpolicy in terms of the breadth of market whichthe company intends to cover. It may choose toconfine its diversifications to the vertical or hori-zontal direction, or it may select a type of lateraldiversification controlled by the characteristics ofthe missions to which the company intends tocater. For example, a company in the field of airtransportation may expand its interest to all formsof transportation of people and cargo. To para-phrase DuPont, some slogan like "Better trans-

portation for better living through advanced en-gineering," would be descriptive of such a long-range policy.

C A gready different policy is to emphasize pri-marily the financial characteristics of the corpo-ration. This method of diversification generaQyplaces no limits on engineering and manufacturingcharacteristics of new products, although in prac-tice the competence and interests of managementwill usually provide some orientation for diversifi-cation moves. The company makes the decisionsregarding the distribution of new acquisitions ex-clusively on the basis of financial considerations.Rather than a manufacturing entity, the corporatecharacter is now one of a "holding company." Topmanagement delegates a large share of its product-planning and administrative functions to the divi-sions and concerns itself largely with coordination,financial problems, and with building up a bal-anced "portfolio of products" within the corporatestructure.

Successful AlternativesThese alternative long-range policies demon-

strate the extremes. No one course is necessar-ily better than the others; management's choicewill rest in large part on its preferences, ob-jectives, skiUs, and training. The aircraft in-dustry illustrates the fact that there is more thanone successful path to diversification:

e Among the major successful airframe manu-facturers, Douglas Aircraft Company, Inc., andBoeing Airplane Company have to date limitedtheir growth to horizontal diversification into mis-siles and new markets for new types of aircraft.Lockheed has carried horizontal diversification fur-ther to include aircraft maintenance, aircraft serv-ice, and production of ground-handling equipment.

C North American Aviation, Incorporated, onthe other hand, appears to have chosen verticaldiversification by establishing its subsidiaries inAtomics International, Autonetics, and Rocketdyne,thus providing a basis for manufacture of com-plete air vehicles of the future.

e Bell Aircraft Corporation has adopted a policyof technological consistency among the items inits product line. It has diversified laterally but pri-marily into types of products for which it had pre-vious know-how and experience.

e General Dynamics Corporation provides a fur-ther interesting contrast. It has gone far into lat-eral diversification. Among the major manufac-turers of air Vehicles, it comes closest to the "hold-ing conipany" extreme. Its airplanes and missilemanufacturing operations in Convair are paralleledby production of submarines in the Electric Boat

Strategies for Diversifcation 121Division; military, industrial, and consumer elec-tronic products in the Stromberg-Carlson Division;electric motors in the Electro Dynamic Division.

Selecting a StrategyIn the preceding sections qualitative criteria

for diversification have been discussed. Howshould management apply th«se criteria to in-dividual opportunities? Two steps should betaken: (i) apply the qualitative standards tonarrow the field of diversification opportunities;(2) apply the numerical criteria to select thepreferred strategy or strategies.

Qualitative EvaluationThe long-range product-market policy is used

as a criterion for the first rough cut in the quali-tative evaluation. It can be used to divide a largefield of opportunities into classes of diversifica-tion moves consistent with the company's basiccharacter. For example, a company whose policyis to compete on the basis of the technical ex-cellence of its products would eliminate as in-consistent classes of consumer products whichare sold on the strength of advertising appealrather than superior quality.

Next, the company can compare each indi-vidual diversification opportunity with the in-dividual diversification objectives. This processtends to eliminate opportunities which, whilestill consistent with the desired product-marketmake-up, are nevertheless likely to lead to animbalance between the company product lineand the probable environment. For example, acompany which wishes to preserve and expandits technical exellence in design of large, highlystressed machines controlled by feedback tech-niques may find consistent product opportunitiesboth inside and outside the industry to which itcaters, but if one of its major diversification ob-jectives is to correct cyclic variations in demandthat are characteristic of the industry, it wouldchoose an opportunity that lies outside.

Each diversification opportunity which hasgone through the two screening steps satisfies atleast one diversification objective, but probablyit will not satisfy all of them. Therefore, beforesubjecting them to the quantitative evaluation,it is necessary to group them into several alterna-tive over-all company product-market strategies,composed of the original strategy and one ormore of the remaining diversification strategies.These alternative over-all strategies should be


roughly equivalent in meeting all of the diver-sification objectives.

At this stage it is particularly important toallow for the unforeseeable contingencies. Sincethe techniques of numerical evaluation are ap-plicable only to trends and foreseeable contin-gencies, it is important to make sure that thedifferent alternatives chosen give the companya broad enough technological base. In practicethis process is less formidable than it may ap-pear. For example, a company in the aircraftindustry has to consider the areas of technologyin which major discoveries are likely to aflEectthe future of the industry. This would includeatomic propulsion, certain areas of electronics,automation of complex processes, and so forth.In designing alternative over-all strategies thecompany would then make sure that each con-tains product entries which will give the firma desirable and comparable degree of participa-tion in these future growth areas.

Quantitative Evaluation

Will the company's product-market strategiesmake money? Will the profit structure improveas a result of their adoption? The purpose ofquantitative evaluation is to compare the profitpotential of the alternatives.

Unfortunately, there is no single yardstickamong those commonly used in business thatgives an accurate measurement of performance.The techniques currently used for measurementof business performance constitute, at best, animprecise art. It is common to measure differ-ent aspects of performance by applying differenttests. Thus, tests of income adequacy measurethe earning ability of the business; tests of debtcoverage and liquidity measure preparednessfor contingencies; the shareholders' positionmeasures attractiveness to investors; tests ofsales efficiency and personnel productivity meas-.ure efficiency in the use of money, physicalassets, and personnel. These tests employ a va-riety of different performance ratios, such as re-turn on sales, return on net worth, return onassets, turnover of net worth, and ratio of assetsto liabilities. The total number of ratios may runas high as 20 in a single case.

In the final evaluation, which immediatelyprecedes a diversification decision, managementwould normally apply all of these tests, tem-pered with business judgment. However, forthe purpose of preliminary elimination of al-ternatives, a single test is frequently used —

return on investment, a ratio between earningsand the capital invested in producing these earn-ings. While the usefulness of return on invest-ment is commonly accepted, there is consider-able room for argument regarding its limitationsand its practical application.^ Fundamentally,the difficulty with the concept is that it fails toprovide an absolute measure of business per-formance applicable to a range of very differentindustries; also, the term "investment" is sub-ject to a variety of interpretations.

But, since our aim is to use the concept as ameasure of relative performance of different di-versification strategies, we need not be con-cerned with its failure to measure absolutevalues. And as long as we are consistent in ourdefinition of investment in alternative coursesof action, the question of terminology is not sotroublesome. We cannot define profit-produc-ing capital in general terms, but we can defineit in each case in the light of particular businesscharacteristics and practices (such as the extentof government-owned assets, depreciation prac-tices, inflationary trends).

For the numerator of our return on invest-ment, we can use net earnings after taxes. Agoing business concern has standard techniquesfor estimating its future earnings. These de-pend on the projected sales volume, tax struc-ture, trends in material and labor costs, produc-tivity, and so forth. If the diversification oppor-tunity being considered is itself a going concern,its profit projections can be used for estimates ofcombined future earnings. If the opportunityis a new venture, its profit estimates should bemade on the basis of the average performancefor the industry.

Changes in Investment Structure

A change in the investment structure of thediversifying company accompanies a diversifica-tion move. The source of investment for thenew venture may be: (i) excess capital, (2) cap-ital borrowed at an attractive rate, (3) .an ex-change of the company's equity for an equity inanother company, or (4) capitd withdrawn frompresent business operations.

If we let ii, ig, is, and i^, respectively, repre-

^ See Charles R. Schwartz, The Return-on-InvestmentConcept as a Tool for Decision Making, General Manage-ment Series No. 183 (New York, American ManagementAssociation, 1956), pp. 42-61; Peter F. Drucker, ThePractice of Management (New York, Harper & Brothers,1954); and Edward M. Barnet, "Showdown in the Mar-ket Place," HBR July-August 1956, p. 85.

sent investments made in the nevp product inthe preceding four categories during the firstyear of diversified operations, we can derive asimple expression for the improvement in returnon investment resulting from diversification:

I + is-l-i

where pi and p2 represent the avera;ge returnon capital invested in the original product andin the nevp product, respectively, and quantityI is the total capital in the business beforediversification.

We can easily check this expression by assum-ing that only one type of new investment will bemade at a time. We can then use the formula tocompute the conditions under which it pays todiversify (that is, conditions where AR is greaterthan zero):

(1) If excess capital is the only source of newinvestment (i2 = i3 = i4 = o), this condition isP2 — r > o. That is, return on diversified opera-tions should he more attractive than current ratesfor capital on the open market.

(2) If only borrowed capital is used (ii = i3 =i4 = o), it pays to diversify if p2 — pi > r. Thatis, the difEerence between return from diversifica-tion and return from the original product shouldbe greater than the interest rate on the money.

(3) If the diversified operation is to be ac-quired through an exchange of equity or throughInternal reallocation of capital, p2 — pi > o is thecondition under which diversification will pay ofE.

A Comprehensive YardstickThe formula for AR just stated is not suffi-

ciently general to serve as a measure of profitpotential. It gives improvement in return forthe first year only and for a particular salestrend. In order to provide a reasonably compre-hensive comparison between alternative over-allcompany strategies, the yardstick for profit po-tential should possess the following properties:

(1) Since changes in the investment structureof the business invariably accompany diversifica-tion, the yardstick should reflect these changes.It should also take explicit account of new capitalbrought into the business and changes in the rateof capital formation resulting from diversification,as well as costs of borrowed capital.

(2) Usually the combined performance of thenew and the old product-market lines is not a sim-ple sum of their separate performances; it should

" See H. Igor Ansoff, A Model for Diversification (Bur-bank, Lockheed Aircraft Corporation, 1957); and John


be greater. The profit potential yardstick musttake account of this nonlinear characteristic.

(3) Each diversification move is characterizedby a transition period during which readjustmentof the company structure to new operating condi-tions takes place. The benefits of a diversificationmove may not be realized fully for some time, sothe measurement of profit potential should span asufficient length of time to allow for effects of thetransition.

(4) Since both profits and investments will bespread over time, the yardstick should use theirpresent value.

(5) Business performance will differ depend-ing on the particular economic-political environ-ment. The profit potential yardstick must some-how average out the probable effect of alternativeenvironments.

(6) The statement of sales objectives, as pointedout previously, should specify the general charac-teristics of growth and stability which are desired.Profit potential functions should be compatiblewith these characteristics.

We can generalize our formula in a waywhich will meet most of the preceding require-ments. The procedure is to write an expressionfor the present value of AR for an arbitrary year,t, allowing for possible yearly diversification in-vestments up to the year t, interest rates, andthe rate of capital formation. Then this presentvalue is averaged over time as well as over thealternative sales forecasts. The procedure isstraightforward (although the alegebra involvedis too cumbersome to be worth reproducinghere*). The result, which is the "average ex-pected present value of AR," takes account ofconditions (i) through (5), above. Let us callit (AR)e. It can be computed using data nor-mally found in business and financial forecasts.

Final Evaluation

This brings us to the final step in the evalua-tion. We have discussed a qualitative methodfor constructing several over-all product-marketstrategies which meet the diversification and thelong-range objectives. We can now compute(AR)e for each of the over-all strategies and, atthe same time, make sure that the strategiessatisfy the sales objectives previously stated, thusfulfilling condition (6), above.

If product-market characteristics, which wehave used to narrow the field of choice and tocompute (AR)e, were the sole criteria, then the

Burr Williams, The Theory of Investment Value (Am-sterdam, The North-Holland Publishing Co., 1938).


strategy with the highest (AR)e would be the"preferred" path to diversification. The advan-tages of a particular product-market opportunity,however, must be balanced against the chancesof business success.

Conclusion

A study of diversification histories shows thata firm usually arrives at a decision to make aparticular move through a multistep process.The planners' first step is to determine the pre-ferred areas for search; the second is to select anumber of diversification opportunities withinthese areas and to subject them to a preliminaryevaluation. They then make a final evaluation,conducted by the top management, leading toselection of a specific step; finally, they workout details and complete the move.

Throughout this process, the company seeksto answer two basic questions: How well will aparticular move, if it is successful, meet thecompany's objectives? What are the company'schances of making it a success? In the earlystages of the program, the major concern is withbusiness strategy. Hence, the first question playsa dominant role. But as the choice narrows,considerations of business ability, of the particu-lar strengths and weaknesses which a companybrings to diversification, shift attention to thesecond question.

This discussion has been devoted primarily toselection of a diversification strategy. We havedealt with what may be called external aspectsof diversification — the relation between a com-pany and its environment. To put it anotherway, we have derived a method for measuringthe profit potential of a diversification strategy,but we have not inquired into the internal fac-tors which determine the ability of a diversifyingcompany to make good this potential. A com-pany planning diversification must consider suchquestions as how the company should organizeto conduct the search for and evaluation ofdiversification opportunities; what method ofbusiness expansion it should employ; and howit should mesh its operations with those of a sub-sidiary. These considerations give rise to a newset of criteria for the lousiness fit of the prospec-tive venture. These must be used in conjunc-tion with (AR)e as computed in the precedingsection to determine which of the over-all prod-uct-market strategies should be selected forimplementation.

Thus, the steps outlined in this article arethe first, though an important, preliminary to adiversification move. Only through further care-ful consideration of probable business successcan a company develop a long-range strategy thatwill enable it to "run twice as fast as that" (usingthe Red Queen's words again) in the ever-chang-ing world of today.

IN a highly diversified company . . . there is a natural tendency toassign a single executive the responsibility for so many diverse businesses

that he becomes a jack of all trades and a master of noneThis is serious, because American business competition no longer per-

mits survival of businesses without managers of special intelligence andcompetence in their individual fields. Therefore, as a continuing process,we attempt to organize our company [W. R. Grace h Co.] so that themanager for any business or group of businesses is as expert in them as hiscompetition. This is sometimes difficult. As one important aid, we havetried to minimize the number of management levels; we have tried to keepthe organization "flat." The more management levels you have, wefeel, the more friction, inertia and slack you have to overcome, and thegreater the distortion of objectives and the misdirection of attention. Inthis you must always be on your guard, because levels of management,like tree rings, grow with age. As one company president put it, "If aU anexecutive does is agree with his subordinate executive, you don't need bothof them."

Ernest C. Arbuckle, "Diversification," Management for Growth, edited byGayton E. Germane

Stanford University, Graduate School of Business, 1957, pp. 85-86.

strategies for diversification

Documents