shirt-sleeve approach to long-range plans
Even if yours is a small cowpany, you can form a better strategy for the next five years if you take a long hard look at what your husiness might be like in twenty years
Robert E. Linneman and ]ohn D. Kennell One bothersome aspect about the future when you're trying to plan is all the uncertainty that surrounds it. In past years forecasters have tried to cope with the problem by coming up with a best guess—a "most probable" scenario—by extrapolating from trends, making a few assumptions, and then suggesting tbat management plan as if tbe best guess were the future. More recently this approacb has fallen into disfavor as recognition has sunk in that, when you sum up all the variables involved which support such a scenario, the odds are very much against it. And tbe consequences for a company can be monumental. Now forecasters are bedging their hets by developing several widely varying scenarios, and urging management to develop a strategy tbat may allow an organization to survive and prosper under any of tbem. But what about the smaller company tbat doesn't have such resources as a forecasting staff and sophisticated
computer hardware and software packages at its disposal? The authors believe that multiplescenario analysis can be of value to its management as well through a less sophisticated "shirtsleeve" approacb, which they describe in this article. Robert E. Linneman is an associate professor of marketing at Temple University. In addition to bis academic work, he has participated in a wide variety of marketing projects in more tban 25 companies and has been director of marketing for a major Philadelphia bank as well as a national director of the American Marketing Association. John D. Kennell is a senior specialist on the operations planning staff at Sun Company, Inc. (formerly Sun Oil Company). His current responsibilities are in the areas of husiness policy analysis and tbe development of methods for strategic issue definition and resolution.
Forecasting and making long-range plans that are based on forecasts are inevitable. And while the importance of forecasting is recognized, so is its main limitation: too many key variables are just too unpredictable. Listen to the laments in 1974 after two humbling years for forecasters: Roderick G. Dederick, chief economist for Chicago's Northern Trust Co.: "We did not provide advance warning to our managements of the distressing situation into which the U.S. economy has drifted over the past several years." Milton W. Hudson, economist for Morgan Guaranty Trust Go.: "We've decided that economic forecasting is so deficient that we're not going to perpetuate any more confusion. There is something misleading about giving people what purport to be accurate numbers when they are really no such thing." Perhaps Paul Samuelson isolated a major problem: "I think that the greatest error in forecasting is not realizing how important are the probabilities of events other than those everyone is agreeing upon." ^ Planners usually choose one "most probable" future environment as the basis of their thinking. They estimate uncontrollable variables as best they can, and a strategy is then developed to achieve the company's objectives. But what recourse does a company have when, because of faulty forecasting, the assumed values of key variables are wrong and its chosen strategy is inappropriate? Even tactical con1. "Two Poor Years for the Forecasters," Business Week, Decemher ai, 1974,
Harvard Business Review
Exhibn I steps in multiple-scenario analysis
March-April 1977
tingency plans may fail to compensate for a faulty strategy. As a consequence, many companies now consider a range of plans that cover several possible environments rather than plans with only one outlook for the future. A partial listing of these companies includes: Dow Chemical, Exxon Corporation, Ford Motor Company, Hewlett-Packard, Marine Midland Bank, Olin Corporation, Sun Company, Uniroyal, and Weyerhaeuser. How-to-do-it literature is scanty, however, and the approaches described usually require a large planning staff, a highly structured long-range planning process, and sometimes even computerized routines. Perhaps the greatest benefits can be gained from a highly sophisticated approach, but we contend that even a "shirt-sleeve" method improves appreciation of the possihilities of the future. Consequently, smaller companies—or even larger companies without extensive planning resources—can devise a more adaptive strategy by using a simplified approach in considering several possihie environments.^ Our purpose in this article is to give a simplified, ten-step approach to developing flexible strategies through what we call multiple-scenario analysis (MSA). Exhibit I illustrates the ten steps in MSA. (Although the process is presented in a step-by-step manner, the loop arrows indicate that it is not necessarily sequential. Multiple-scenario analysis involves many experiments that usually necessitate backtracking.)
Simplified analysis This ten-step procedure assumes that a person, or perhaps a committee, can devote only part of his (or its) time to long-range planning. After descrihing each step, we shall use a case example of corporate-level planning at a hypothetical company named "Quik-Serv" to show how the step would be carried out.
Step one Identify and make explicit your company's mission, basic objectives, and policies.
Long-range plans
At least on the first attempt, assume that your company's mission, objectives, and policies won't change. In order to estahlish a base for the following steps, you need to state these explicitly. Mission: What is your company's hasic reason for heing in husiness? Prohahly hecause of existing strengths and commitments, the mission is unlikely to change over the next several years. Consequently, only after present husiness potential has heen carefully examined and found wanting should other possihilities he examined.
this horizon. The purpose of MSA is not to enable you to improve detail and precision in planning farther into the future. Rather, it is to give you an improved appreciation of the possihie variations in future environments in which you must operate and, suhsequently, the long-range implications of today's decisions.
For example, short-range forecasting might indicate a sales decline in the next three years. Given this forecast, one might follow a strategy of across-theboard cutbacks. On the other hand, hecause of discontinuities, certain scenarios might depict considerBasic objectives: Does your company have long-term ahly higher sales in ten years (see Exhibit II]. Such ohjectives for such factors as return on investment, long-range forecasts also would probably call for earnings per share, or size? If explicit minimum retrenchment, but on a more selective basis to mainacceptable standards have not heen set, what are tain a growth posture. In this respect, consideration common-sense minimums? of several scenarios makes the implications of present decisions more apparent. Policies: Are there certain limitations, either explicit or implicit, such as "will not expand overseas," The time length for scenarios is arbitrary, but genor "must provide johs for existing employees?" erally it should he at least five years. The acid test is, of course, "How far in the future are you comExplicit planning premises mitting your resources?" Quik-Serv, our hypothetical company, is a small private-hrand gasoline marketer. The company retails Quik-Serv's planning horizon is 5 years. While gasoline under its own hrand name through a netmanagement has little confidence in its ahility to work of company- and dealer-operated outlets. Quikplan this far in advance, in some respects a 5-year Serv is nonintegrated, buying its products from planning horizon is too short. Quik-Serv's evaluamajor refiners. tion and depreciation of capital investments, for instance, is based on a 20-year economic life. ThereQuik-Serv's mission is to serve consumer needs for fore, management decides to develop a set of sceautomotive fuels. Its objectives are to remain comnarios with a 15- to 20-year horizon. petitive in the marketplace of the future, and to increase aftertax return on stockholders' equity to 12% by 1982. It has two policies: (1) the company's Step thiee behavior will he both legal and ethical, and (2) new lines of business must be related to the company's Develop a good understanding of your company's mission and support its ohjectives. points of leverage and vulnerability.
Step two Determine how far into the future you wish to plan. Your current planning horizon is probably one, two, or perhaps as many as five years. Time and planning facility limitations may prevent you from extending 2. Some of the most helpful literature follows: Rene D. Zentner, "Scetiatios: A New Tool for Corporate Planners," Chemical and Enpneszing News, Industrial Edition, October 6, 1975, p, 22; Ian H. Wilson, "Futures Forecasting for Strategic Planning at Genetal Electrie," Long Range Planning, June 1973. P- 39; Frank L. Mordaod, "Dialectic MetJiods in Forecasting," The Fatmist, August r97!, p. 169, Peter F. Chapman, "A Method of Exploring the Future," Long Range Planning, February 1976, p. 2, and MJ. Creton and Audrey Clayton, "Social Forecasting: A Practical Approach," in The Nsxt 2j Years,- Crisis and Opporttmify, ed. Andrew A. Speklte (Washington, D.C.: Woild Future Society, 1,75!, p. 267.
First, take a look at your industry. What was it like a decade or two ago? What is it like today? What are the causes for the changes? Next, examine your own company. Look hack over the same time span that you did for your industry. Then, take a look at your company today. Given conditions in the industry, what are the similiarities and/or differences hetween your industry and your company? What are your points of leverage and vulnerability? Although these analyses should he in writing, they need not be detailed. You are only seeking basic understandings.
Harvard Business Review
March-April 1977
Exhibit il Trend iine for saies industry Estimatad range of industry sales in scenario #2 Estimated range of industry sales \ in scenario #11
[ Forecasted saies
Be careful of myths. In any husiness enterprise there is an ahundance of information and opinion; unfortunately, much of it is false. From the history of Sears, Roehuck and Go., Peter Drucker came to the following conclusion: "The right answers are always obvious in retrospect. The basic lesson of the Sears story is that the right answers are likely to he anything hut obvious before they have proven themselves. 'Everyhody knew' around 1900 that to promise 'satisfaction guaranteed or your money back' could only bring financial disaster to a retailer. 'Everybody knew' around 1925 that the American market was sharply segmented into distinct income groups.... 'Everyhody knew'— as late as 1950—that the American consumer wanted to shop downtown, and so on." ' Test for myths. Get data. Summarize past trends. Do they reinforce or oppose what you thought you knew? Look at past and present
A decade ago in Quik-Serv's industry, fuel was cheap and abundant, and demand was growing steadily at approximately 4% a year. The major brand fullservice station was the most prevalent type of outlet, and the numher was increasing. Volume, not profitability, was the industry-wide criterion of success. Also, a pump-price spread of several cents existed between major and independent brands.
Today, the situation is almost exactly the opposite. Fuel is considerably more expensive, and demand is leveling off, with only a 2% near-term annual growth rate projected. The full-service outlet is still the most prevalent type hut is losing ground to gasoline-only and self-service outlets. With profitahility as the criterion of success, the outlet population is shrinking as companies close unprofitable units and withdraw from entire geographical areas. Finally, the pump-price spread between majors and independents is becoming blurred as majors increasingly offer self-service islands and gasoline-only outlets. As for the company itself, ten years ago Quik-Serv was a small regional marketer with just over 100 outlets. Competing on a price basis, Quik-Serv enjoyed a 5% to 6% share of each local market it served. The company's aftertax return on stockholders' equity during this period of severe price competition and low margins averaged 6% to 7%. QuikServ's aggressive pricing strategy caused its gasoline volume to grow at ahout 6% per year, a rate some 50% higher than the industry average. The company's 100-odd outlets were of the gasoline-only type, some having been converted from acquired full-service outlets and others of newer construction without service hays. After ten years of growth, Quik-Serv today has 16s modern gasoline-only outlets. Industry-wide conditions of a more stahle pricing structure and wider
Long-range plans
margins have raised return on equity to just over 9%. Rapid inflation and a high cost of capital have significantly slowed Quik-Serv's expansion program. Volume at existing outlets is growing at about the same rate as the market. Quik-Serv's management helieves the company has two major competitive advantages: a lean organization with capable, motivated management, and a chain of modern, efficient outlets in good markets and good locations. Management has also identified two points of vulnerability: dependence on other companies for product supply, and the fact that the company serves a highly price-conscious market segment which has little brand loyalty.
Step four Deteimine factors tbat you think will definitely occur within youi planning time frame. Some assumptions might stem from certain factors that ean be forecasted with almost complete certainty (in Quik-Serv's case, for example, the number of licensed drivers in 1997), and you might consider some information, such as calculated natural resources, accurate and conclusive. Qther projections, such as rate of economic growth, are unpredictable and should he classified as variables. Assumptions must be accurate and conclusive, not merely variables with a high probability of occurrence. A test: if a "prudent man" could doubt its value, consider it a variable. Quik-Serv made the following assumptions: (a) by 1997 there will be a 9% increase in licensed drivers; (b) gasoline will remain the major fuel for automobiles; (e) govemment polieies will be protective of small independent eompanieS; (d| there will be less gasoline brand loyalty; (e) mass merehandisers will eontinue to increase in importance in tires, batteries, accessories (TBA), and repairs; and (f) improved technology will produce autos and TBA requiring less maintenance and repairs, and better gas mileage.
Your main consideration at this step is to identify, without going into a lot of detail, the variables that have been crucial to your company in the past and those that will be important to it in the future. Try to use key variables that are commonly predicted and monitored, such as GNP and the rate of inflation. Easily identifiable vital signs facilitate forecasting and simplify control. If time and resources permit, you may wish to broaden your viewpoint by scanning future-oriented periodicals, sueh as Futures and The Futurist, in addition to your normal fare of business and general periodicals. Of course, there are also a number of proprietary services available, such as the Arthur D. Little Impact Service, the National Planning Assoeiation Eeonomic Projections Series, Predieasts, the Futures Group Scout Service, and the Stanford Research Institute Business Intelligence Program. You can keep the planning task on a "shirt-sleeve" basis by limiting key variables to no more than four or flve by using the following guidelines: D
Delete variables having a low probability of occurrence and a low potential impact. On the other hand, include those with low probabilities but high impaet and all those with high probabilities regardless of their impact. D Consider the timeliness of the variable. Because the future is so unpredietable, it is more important to include an event that is likely to happen or have an impact in the next few years than one that may not happen or be insignifieant until near the end of the planning horizon.
n
Delete disaster events. Events that would eause total disaster—such as a major nuclear war—should not be eonsidered seriously. D Aggregate when possible. For example, the factors responsible for economic growth include, to name a few, expenditures on consumer, investment, and government goods. If only the economic growth rate is relevant, just use it as the representative variable for your analysis. D Step five Separate dependent from independent variables. Check for interdependenee. Is the value of one variMake a list of key variables that will have make-or- able based upon the value of another? If so, then bieak consequences for youi company. remove the dependent variable. Keep a separate list of dependent variables for use in building and en3. Peter F. Diucker, Manflgament; ' riching scenarios. York: Harper & Row, 1974), p. 57.
Harvard Business Review
ables, dependent variables, and assumptions,- and (3) developing a narration describing the future under this set of conditions.
Exiiibit iti Variable value matrix Variable
Value(15to20yef irs in the future) Extreme
Middle.
Extreme
Inflation rate
10% to 20%
-
5% to 9%
Price and avajiability of gasoline
Same as present —
Growth in GNP
More than 4%
1.5% to 4%
March-April 1977
Tight supply high prices Less than 1.5%
Quik-Serv thinks that the rate of inflation, GNP (as an indicator of car sales and usage), and gasoline price and availability are the variahles that will have the greatest impact on its operations. The company's list of dependent variables are: availability of capital |as affected by rate of inflation and growth rate of GNP), government taxation (taxes to curtail consumption), government restriction (rationing), and the rate of technological change (as affected by growth rate of GNP).
When building scenarios, the following suggestions may be helpful.
n Develop at least three, but no more than four, scenarios. Experience has shown that two are too few; the scenarios tend to be classified as good and bad. Qn the other hand, even four scenarios may be too many unless tbey bave markedly different characteristics.* Qne of the scenarios should be for the most probable case.
•
Select values of key variables so that each scenario is distinct from the others. Scenarios with small variations have little value since strategies for these scenarios would be almost identical. In fact, it might be wise to look at a "deadly enemy" scenario.' D Keep the scenarios plausihle. This is partially accomplished by ineluding only the values tbat are deemed realistie. But make sure tbat the comhined Step six variable mix in each scenario also makes sense, that it is feasible. For example, low double-digit inflation, Assign reasonable values to each key variable. rapid economic growth, and severe oil shortage by themselves might seem plausible but their joint ocNow you need to pick a reasonable range over which each variable may vary, and divide tbe range into currence does not. To keep from overlooking plausible combinations of key variables, examine all two or three sets of values—a "middle ground" and possibilities. the extremes. Reasonability, of course, can only be D determined by common sense. However, two helpful principles are: (1) reject values so extreme that they In writing a scenario, first state variables and assumptions in an abbreviated form. Then include seem absurd^ and (2) if a value lies between the the dependent variables and develop the seenario marginal and the absurd, use it. Although it is generally recommended that three sets of values be with more description. Write from the viewpoint of someone standing in the future (at the end of estimated, in some instances two may suffice for all your time frame) descrihing conditions at that time practical purposes. and how they developed. The completed scenario should transmit an appreciation of this hypothesized To maintain objectivity, you may want to seek the environment. opinions of fellow executives, trade association officials and staff, or, if relationships permit, customers D Limit the length of each scenario to one or two paraand suppliers. The ranges of values for the key variables that Quik-Serv picked are sbown in Exhibit III. graphs, keeping the length of eaeh scenario the same, and using common language and classifications to allow for point-hy-point comparison. D Step seven Keep the themes of eaeh seenario neutral. Although Build scenarios in which your company may operate. you may have scenarios for the "worst," "most probable," and "best" circumstances, avoid labeling Scenarios describe possible future operating environ- them as such to prevent more consideration being ments for your organization. A scenario is built by given to the "most probable" or "best" seenarios. (1) seleeting a value for eaeh key variable; (2) esti- 4 Rene S. Zentner, "Scenarios: A New Tool." mating the resulting interactions between key vari- S. Frank L. Moreland, "Dialectical Methods."
i
Long-range plans
Exhibit IV Matrix for testing plausibility
Possible scenario
Inflation
Gas price and availability
GNP
10% to 20%
Present
More than 4%
10 to 20
Tight supply, high prices
More than 4
10 to 20
Present
1.5 to4
10 to 20
Tight supply,-high prices
1.6 to 4
Not piausibie
Notpiausibie
Tight supply, high prices 5to9
Present
More than 4
5to9
Tight supply, high prices
More than 4
9
5to9
Present
1.5 to 4
10
5 to 9
Tight supply, high prices
1.5 to 4
5 to 9
Present
Less than 1.5
Not plausible
Not piausibie
5 to 9
Quik-Serv's scenarios
Exhibit IV shows the matrix Quik-Serv constructed to test the plausibihty of the comhined variables. The 12 scenarios hsted represent all possihie combinations of the three key variables. The company rejected Scenarios 2, 4, 8, and 10 because of inconsistencies among the major variables. Scenarios 3, 6, 9, and 11 were selected for further development. Number 9 was thought to represent the middle of the variables' possible ranges, while numbers 3, 6, and 11 were chosen to represent more extreme possibilities. Scenarios 1, 5, 7, and 12 were not chosen because they were thought to be merely variations or extreme cases of the general themes of Scenarios 3, 6, 9, and 11 respectively. (For the sake of brevity, we shall only consider Scenarios 3 and 11 further.) Scenario 3: Advances in technology and design have made the automobile of the early 1990s safer, more reliable, and nearly service-free. From the viewpoint of the gasoline retailer, however, the most significant change in the automobile has been the tremendous increase in fuel efficiency over the past two decades. Better mileage has offset the effects of increases in the number of cars on the road. Overall market demand has grown at an average of only 1% to 1.5% per year. Because of a low rate of market growth and a high cost of capital, the number of retail gasoline outlets has increased only slightly since the late 1970s.
Rapidly escalating labor costs and the "service-free" automohile have been responsible for the demise of the full-service gasoline outlet. The high rate of inflation has made price-conscious consumers extremely receptive to the economies of self-service stations. The price competition between majors and independents is head-on, so there is no pump-price spread. Scenario 11: The gasoline-powered automobile is still the dominant mode of transportation. Yet, gasoline demand, which peaked in the late 1980s, has been declining slightly each year during this period of prolonged economic stagnation. Gasoline outlets have been decreasing in number since the early 1980s as retailers attempt to cover rising fixed costs in the face of static demand and strong consumer resistance to price increases. Low levels of profitability no long:er justify the use of new capital in the retail gasoline business. One important countertrend has been the growth of the full-service outlet as hard-pressed consumers increasingly choose to repair rather than replace their automobiles. Marketers in a position to do so strive to retain brand identification by building consumer loyalty through the service and repair portions of their operations. The economics of the full service outlet benefit from the general availability and the relatively low cost of labor. The federal government, concerned with the survival of the
148
Harvard Business Review
small companies in these difficult times, is especially watchful for anticompetitive practices by the larger companies.
Exhibit V Flexibility of strategies
strategies
Step eight
Maich-April 1977
Fit in Scenario 3
Finn Scenario 11
Scenario 3
Develop a strategy for each scenario that will most likely result in achieving your company's objectives.
Ownership of outlets
-
Poor
Increase debt posture
-
Poor
The procedure for generating strategies is similar to traditional strategy development. Because of the "unpredictable future," however, a greater emphasis should be placed on flexible strategies that have relatively high payoffs in short and intermediate time spans.
Selective expansion
-
Dealer-operated outlets
-
Fair
Automated self-service
-
Fair
Your "final" strategies, within their respective scenarios, should enable your company to come reasonably close to its objectives. Because of the imprecision inherent in long-term forecasting, avoid the "numbers game." Instead, make common-sense, intuitive judgments. Consttained opportunity...
The overall theme of Quik-Serv's Scenario 3 is one of constrained opportunity. That is, there is some opportunity for growth in the market segment in which Quik-Serv believes it has competitive advantages (gasoline-only), but this opportunity is constrained by the rather low rate of market growth, the high rate of inflation, the cost of capital, the high cost of labor, and head-on price competition from the majors. With this long-term view of the future, Quik-Serv's management formulates the following selective growth strategy: Ownership of outlets: Strive for ownership of outlets and real estate on a leveraged basis as a hedge against inflation. Increased debt posture: Increase debt/equity ratio as necessary to facilitate buying and building now instead of later during the inflationary period.
.
Poor
Scmarloii Restricted ownership
Fair
-
Minimize debt
Poor
-
Selective divestment
Fair
-
Company-operated outlets
Fair
-
Increase labor
Poor
-
Automated self-service: Further minimize the effect of escalating labor costs at remaining companyoperated outlets by making a shift to automated self-service. . . . vs. recessionary outlook
The message emerging from Scenario 11 is threefold: a decline in the demand for gasoline, consumer resistance to price increases, and a decreasing number of service stations. Quik-Serv's management decides that under this set of conditions, the following strategy is the most feasible course of action: Restricted ownership of outlets: Reduce ownership of low and negative profit outlets where real estate is usable as gasoline outlet only. Increase ownership of properties which have attractive alternatives for commercial usage. Minimize debt: Avoid new debt and use excess eash to retire existing high-interest obligations.
Selective expansion: Maintain share of market growth through highly selective and carefully timed expansion with new gasoline-only outlets, giving careful consideration to profitability criteria and market supply-demand balances.
Selective divestment: Identify and divest outlets which make zero or negative contribution to profit and for which turnaround would require large amounts of new capital.
Dealer-operated units: Minimize the effect of escalating labor costs by increasing emphasis on dealeroperated outlets and decreasing operation of outlets by direct company employees.
Company-operated outlets: Increase profitability by stressing company operations, thus absorbing dealer margins during this period of readily available and relatively low-cost labor.
Long-range plans
Increase labor: Utilize inexpensive labor rather than strategy outlined for Scenario 3; with three modifiautomated self-service equipment which is relacations to improve its Hexihility: tively eostly. Go slow on shift to dealer operations, sinee, if necessary, it would he easier to shift operations from the Step nine company to dealers rather than to shift from dealers back to the eompany. Check the flexibility of each strategy in each scenario by testing its effectiveness in the other scenarios. Increase debt/equity ratio slowly by implementing a more selective expansion program. Is each strategy adaptahle to your other scenarios, or does its effectiveness depend on the values of key Prepare for a selective divestment strategy by identivariahles in the particular scenario for which it was fying a list of candidates. originally developed? The construction of a matrix and subsequent ratings, as shown in Exhibit V for Multiple-seenario analysis gave Quik-Serv manageQuik-Serv, will help you to visualize each strategy's ment a better appreciation of the long-term outlook, adaptability. thus providing an improved foundation for its strategic five-year planning.
Step ten Select—or develop—an "optimum response" strategy. Now that you have developed a strategy for each scenario, you must choose one of them or form a "compromise" strategy from among them. Beeause of personal and eompany attitudes toward risk, there can be no hard-and-fast rules to apply as you make this ehoiee. However, in general, the final strategy should: D
Provide maximum adaptability to the conditions of the several scenarios, or conversely, require only a short reaction time for adjustment to the demands of different environments. D Have favorable consequences in scenarios with relatively high probabilities of occurrence.
n Be partieularly attractive in the near future since the distant future is less predictahle. D
Provide for maximum delay of expenditures, taking the impact of possible inflationary price rises into eonsideration. Quik-Serv's management believed a future environment similar to Scenario 3 was highly probable, especially over the next s to 7 years. It did not totally diseount the harsher possibilities of Scenario 11, but felt that if sueh a condition did materialize it would not be until the latter part of the 15- to 20year period. Accordingly, it deeided to pursue the
Pitfalls & benefits of eourse, there may be a gap between your objeetives and common-sense estimates of what you can hope to achieve with the "optimum response" strategy. If so, you should first try to shrink the gap to tolerable limits by developing a better strategy. Perhaps in the planning proeess some factors have been overlooked or over- or underemphasized. Consider bringing in other viewpoints. Only afterwards should new ventures be investigated to supplement, or possibly replace, the basic mission. To analyze new ventures, first list potential candidates. Seek those highly compatible with the basic mission. Fast screening of this list is necessary because, as a rule, new ventures will require separate sets of scenarios sinee some of their key variables will be different from those of the basic mission. Identify and eliminate weak eandidates by working through, for each potential new venture, steps four to ten on a eursory basis. Then, perform a more detailed analysis for eaeh of the more plausihle ventures. Although some of the make-or-break variables for venture eandidates may be different from those of the basie mission, some will probably be the same (for example, the rate of inflation|. Use, when possible, identical values for the same key variables in building scenarios. Then, in seleeting the final strat-
Harvard Business Review
egy (for both the basic mission and the new ventures), make sure comparable scenarios are used.
Taking corrective action A major concern is "Will the actual environment, as it unfolds, provide a favorable climate for the implemented strategy?" Consequently, there must be an environmental monitoring system. To develop such a procedure, in general, first establish shortterm standards for key variables that will tend to validate the long-range estimates. Although favorable long-range values have been estimated, shortterm guidelines are needed to indicate if the scenario is unfolding as hoped. Next, set up criteria to decide when the strategy must be changed. Of course your decision will depend on the magnitude or the trend (or both) of the deviations. If the optimum strategy was developed for environmental conditions different from those that actually are occurring, then examine how effective the strategy was judged to be in a scenario closer to the actual conditions, and determine necessary adaptions. In this manner your analysis of different scenarios also can be used to respond to "unforecasted" situations as they develop.
Common sense Multiple-scenario analysis does improve the appreciation of the future, and so should enable management to develop strategies with better cognizance of potential risks. At first, however, the procedure may appear to "muddy the water," since it calls for the development of more than one possible environment, and more than one strategy. Also, you need to be aware of problems that might be caused by misusing MSA. At the beginning of the planning process, for instance, you must carefully avoid favoring what may seem to be the right scenario or the right strategy. Such prejudice minimizes the value of the process, as it would any type of planning model. Also, be careful of adopting a strategy simply because it fits more scenarios than does any alternative. It might not be the strategy most likely to lead to success. Finally, in spite of your best laid plans, you may have missed developing a scenario that portrays an environment close to the one that actually occurs. Hence, it is necessary to adopt a systematic method
March-April 1977
of monitoring the environment and to consider long-range planning as a continuous process. If you are aware of these pitfalls, the process can help you to focus on the assumptions underlying the scenarios, not on just whether to approve a plan or not. Actually, MSA is simply a matter of common sense. It is essentially a part of a decision-making procedure that we all use daily, although on a highly informal basis. By formalizing the process, however, you facilitate communication, thus improving multiparty participation and comprehension. And there is an important side benefit: a structured approach leads to more rigorous thinking. As a result, multiple-scenario analysis can be a valuable tool in helping to make today's decisions fiexible enough for the uncertain future.
