The Principles of Decision Making

The need for speed places a premium on efficient decision making. But effective strategic moves also require organizational buy-in, which is best achieved through time-consuming consensus building. Harvard Business School assistant professor of general management Michael Roberto reveals some mechanisms for achieving both in his working paper, “Strategic Decision-Making Processes: Beyond the Efficiency-Consensus Tradeoff.” Broad-based consensus on decisions to launch a new product line, for example, or to pursue a major contract, facilitates the resulting implementation of the strategy. But a trade-off emerges because along with more seats at the table comes more time required to build consensus, and a greater risk of the analysis paralysis that typifies large organizations. Indeed, scholars have long noted the difficulties a trade-off presents for companies and policymakers alike.

Roberto describes a set of do's and don'ts to maximize chances of attaining both efficiency and consensus emerging from his 1999-2000 field study of 10 major decisions at a respected aerospace company. First, the do's. “[Good groups] embark on a gradual structuring process instead of putting all the options on the table and trying to make a decision,” says Roberto. “Through a series of intermediate agreements, you can manage conflict and debate effectively.” The paper describes three strategies: having a clear decision criteria defined early on; using a winnowing process to narrow down a set of alternatives; and determining contingent options, that is, defining criteria by which to eliminate options as the process unfolds.

Such a rational approach seems foolproof, but pitfalls remain. Even the best-structured analysis can be neutralized by a poor political approach that delegit-imizes the process. That's where the don'ts figure in. “The poorly performing groups want to make the process look legitimate,” says Roberto, by making an analysis seem comprehensive when it is not. But when steps are perceived as political rather than analytical, it subverts the perceived role of participants. And that undermines buy-in, one reason for bringing disparate individuals to the table in the first place. The danger is when, as Roberto points out, “you do them purely for legitimacy reasons, and they aren't authentic. Then you harm both speed and consensus,” because others see through the artifice.

For example, it's advantageous to analyze a number of alternatives. But if some alternatives are presented as tokens — they are discussed solely to increase the number of options on the table — then others perceive the process as flawed. Similarly, one requirement for rigorous analysis is to gather much information. But some groups then share the information unequally. They provide selected colleagues with more information than others, in an attempt to build coalitions. That political behavior damages the credibility of the process in the eyes of those spurned.

Finally, a comprehensive process implies a formal analysis. Indeed, successful groups all used some form of formal analysis. But the analysis must appear to be unbiased; when it is driven by strong advocates of a particular plan, again, the process loses credibility. Managers sense that other alternatives were never under consideration.

The same do's and don'ts might not be valid for all companies. “I don't think [they're] applicable to small firms,” Roberto notes, because they tend to be flatter organizations with influential founders. Consensus, presumably, is less of an issue in those businesses. In addition, the aerospace company in the study was led by analytically minded, financial types. The results might differ at, say, an advertising firm dominated by creative types.

The paper, which originated as Roberto's doctoral thesis and won best doctoral student prize in the business policy unit of the Academy of Management in 2000, was reviewed by and is being revised for Organization Science. Roberto interviewed and surveyed 78 employees at various levels in the organization, and reviewed both internal and external documents. Three of the 10 decisions he studied, incidentally, were classified as high efficiency and high consensus. And, yes, they proved to be better decisions than others the company made.