The traditional approach to strategy requires precise predictions and thus often leads executives to underestimate uncertainty. This can be downright dangerous. A four-level framework can help.

At the heart of the traditional approach to strategy lies the assumption that executives, by applying a set of powerful analytic tools, can predict the future of any business accurately enough to choose a clear strategic direction for it. The process often involves underestimating uncertainty in order to lay out a vision of future events sufficiently precise to be captured in a discounted-cash-flow (DCF) analysis. When the future is truly uncertain, this approach is at best marginally helpful and at worst downright dangerous: underestimating uncertainty can lead to strategies that neither defend a company against the threats nor take advantage of the opportunities that higher levels of uncertainty provide. Another danger lies at the other extreme: if managers can’t find a strategy that works under traditional analysis, they may abandon the analytical rigor of their planning process altogether and base their decisions on gut instinct.

Making systematically sound strategic decisions under uncertainty requires an approach that avoids this dangerous binary view. Rarely do managers know absolutely nothing of strategic importance, even in the most uncertain environments. What follows is a framework for determining the level of uncertainty surrounding strategic decisions and for tailoring strategy to that uncertainty.

Four levels of uncertainty

Available strategically relevant information tends to fall into two categories. First, it is often possible to identify clear trends, such as market demographics, that can help define potential demand for a company’s future products or services. Second, if the right analyses are performed, many factors that are currently unknown to a company’s management are in fact knowable—for instance, performance attributes for current technologies, the elasticity of demand for certain stable categories of products, and competitors’ plans to expand capacity.

The uncertainty that remains after the best possible analysis has been undertaken is what we call residual uncertainty—for example, the outcome of an ongoing regulatory debate or the performance attributes of a technology still in development. But quite a bit can often be known despite this. In practice, we have found that the residual uncertainty facing most strategic-decision makers falls into one of four broad levels (Exhibit 1).

Level one: A clear enough future

The residual uncertainty is irrelevant to making strategic decisions at level one, so managers can develop a single forecast that is a sufficiently precise basis for their strategies. To help generate this usefully precise prediction of the future, managers can use the standard strategy tool kit: market research, analyses of competitors’ costs and capacity, value chain analysis, Michael Porter’s five-forces framework, and so on. A DCF model that incorporates those predictions can then be used to determine the value of alternative strategies.

Level two: Alternative futures

The future can be described as one of a few discrete scenarios at level two. Analysis can’t identify which outcome will actually come to pass, though it may help establish probabilities. Most important, some, if not all, elements of the strategy would change if the outcome were predictable.

Many businesses facing major regulatory or legislative change confront level two uncertainty. Consider US long-distance telephone providers in late 1995, as they began developing strategies for entering local telephone markets. Legislation that would fundamentally deregulate the industry was pending in Congress, and the broad form that new regulations would take was fairly clear to most industry observers. But whether the legislation was going to pass and how quickly it would be implemented if it did were still uncertain. No amount of analysis would allow the long-distance carriers to predict those outcomes, and the correct course of action—for example, the timing of investments in network infrastructure—depended on which one materialized.

In another common level two situation, the value of a strategy depends mainly on competitors’ strategies, which cannot yet be observed or predicted. For example, in oligopoly markets, such as those for pulp and paper, chemicals, and basic raw materials, the primary uncertainty is often competitors’ plans for expanding capacity. Economies of scale often dictate that any plant built would be quite large and would be likely to have a significant impact on industry prices and profitability. Therefore, any one company’s decision to build a plant is often contingent on competitors’ decisions. This is a classic level two situation: the possible outcomes are discrete and clear, and it is difficult to predict which will occur. The best strategy depends on which one does.

Here, managers must develop a set of discrete scenarios based on their understanding of how the key residual uncertainties might play out. Each scenario may require a different valuation model. Getting information that helps establish the relative probabilities of the alternative outcomes should be a high priority. After establishing an appropriate valuation model for—and determining the probability of—each possible outcome, the risks and returns of alternative strategies can be evaluated with a classic decision analysis framework. Particular attention should be paid to the likely paths the industry might take to reach the alternative futures, so that the company can determine which possible trigger points to monitor closely.

Level three: A range of futures

A range of potential futures can be identified at level three. A limited number of key variables define that range, but the actual outcome may lie anywhere within it. There are no natural discrete scenarios. As in level two, some, and possibly all, elements of the strategy would change if the outcome were predictable.

Companies in emerging industries or entering new geographic markets often face level three uncertainty. Consider a European consumer goods company deciding whether to introduce its products to the Indian market. The best possible market research might identify only a broad range of potential customer penetration rates—say, from 10 percent to 30 percent—and there would be no obvious scenarios within that range, making it very difficult to determine the level of latent demand. Analogous problems exist for companies in technologically driven fields, such as the semiconductor industry. When deciding whether to invest in a new technology, producers can often estimate only a broad range of potential cost and performance attributes for it, and the overall profitability of the investment depends on those attributes.

The analysis in level three is similar to that in level two: a set of scenarios describing alternative future outcomes must be identified, and analysis should focus on the trigger events indicating that the market is moving toward one or another scenario. Developing a meaningful set of scenarios, however, is less straightforward in level three. Scenarios that describe the extreme points in the range of possible outcomes are often relatively easy to develop but rarely provide much concrete guidance for current strategic decisions. Since there are no other natural discrete scenarios in level three, deciding which possible outcomes should be fully developed into alternative scenarios is a real art. But there are a few general rules. First, develop only a limited number of alternative scenarios—the complexity of juggling more than four or five tends to hinder decision making. Second, avoid developing redundant scenarios that have no unique implications for strategic decision making. Third, develop a set of scenarios that collectively account for the probable range of future outcomes and not necessarily the entire possible range. Establishing the range of scenarios should allow managers to decide how robust their strategies are, to identify likely winners and losers, and to determine, at least roughly, the risk of following status quo strategies.

Level four: True ambiguity

A number of dimensions of uncertainty interact to create an environment that is virtually impossible to predict at level four. In contrast to level three situations, it is impossible to identify a range of potential outcomes, let alone scenarios within a range. It might not even be possible to identify, much less predict, all the relevant variables that will define the future.

Level four situations are quite rare, and they tend to migrate toward one of the others over time. Nevertheless, they do exist. Consider a telecommunications company deciding where and how to compete in the emerging consumer multimedia market. The company will confront a number of uncertainties concerning technology, demand, and relations between hardware and content providers. All of these uncertainties may interact in ways so unpredictable that no plausible range of scenarios can be identified.

Companies considering major investments in postcommunist Russia in 1992 faced level four uncertainty. They could not predict the laws or regulations that would govern property rights and transactions—a central uncertainty compounded by additional uncertainty about the viability of supply chains and about the demand for previously unavailable consumer goods and services. Shocks such as a political assassination or a currency default could have spun the whole system toward completely unforeseen outcomes.

This example illustrates how difficult it can be to make strategic decisions at level four but also underscores the transitory nature of level four situations. Greater political and regulatory stability has turned decisions about whether to enter Russian markets into level three problems for most industries today. Similarly, uncertainty about strategic decisions in the consumer multi- media market will migrate to level three or to level two as the industry begins to take shape over the next several years.

Situation analysis at level four is highly qualitative. Still, it is critical to avoid the urge to throw up your hands and act purely on instinct. Instead, managers need to catalog systematically what they know and what it is possible to know. Even if it is impossible to develop a meaningful set of probable, or even possible, outcomes, managers can gain a valuable strategic perspective. Usually, they can identify at least a subset of the variables determining how the market will evolve over time. They can also identify favorable and unfavorable indicators of these variables—indicators that will let them track the market’s evolution over time and adapt their strategy as new information becomes available. By studying how analogous markets developed in other level four situations, by determining the key attributes of the winners and losers, and by identifying the strategies they employed, managers can also identify patterns that show how the market may evolve. Finally, although it will be impossible to quantify the risks and returns of different strategies, managers should be able to identify what information about the future they must believe to justify the investments they are considering. Early market indicators and analogies from similar markets will help sort out whether such beliefs are realistic (see sidebar, “Postures and moves”).