The Business of Managing Decisions

If introducing a new product line creates a small probability of losing a few million dollars and a large probability of earning many billions of dollars, it seems clear that the new product line should be introduced. But does what we know about managerial decision making suggest that the new product line will be introduced?

Unfortunately, a great deal of research suggests that many decisions involving uncertainty go awry. It appears that people make systematic errors in reacting to the probabilities of events. In particular, decision makers may exaggerate the importance of small probabilities, understate the importance of large probabilities, and are generally insensitive to changes in probability that fall between those two extremes.

In recent research, "Money, Kisses and Electric Shocks: On the Affective Psychology of Risk," Christopher Hsee, professor of behavioral science at the University of Chicago Graduate School of Business, and Yuval Rottenstreich, assistant professor at the school, found that this pattern of exaggeration, understatement, and intermediate insensitivity is especially pronounced when the outcomes of a decision are emotionally charged -- as they so often are in crucial business decisions.

An exaggeration of the small probability of a loss and an understatement of the large probability of a gain may inappropriately discourage a manager from making an important decision, such as launching a new product line.

What's to be done? As a first step, organizations need to recognize when emotions influence decisions. There are two key steps in the managerial decision making process:

• Managers must thoroughly research and accurately assess the likelihood of various possibilities.
• Managers must properly and expertly apply their knowledge.

Then, the simplest remedy, says Rottenstreich, is to separate the two steps of decision making. One manager should be charged with task of assessing the likelihood of success and failure. Another manager should be responsible for making decisions based on the given assessments.

"This sort of separation makes it more likely that the person who makes the final decision will be emotionally detached from the outcome and will react to the probabilities of the relevant possibilities," says Rottenstreich. Only then can one ensure that decisions are made in a more rational and calculated way.

The Buck Stops Here

Errors in reacting to the probability of an event can be explained as follows. One famous study found that the typical person was indifferent between receiving a 1 percent chance of winning $200 and receiving $10 without doubt and was also indifferent between receiving a 99 percent chance of winning $200 and receiving $188 for certain.

That is, people reacted to probabilities in a way that made the first hundredth of probability worth $10, the last hundredth worth $12, but the ninety-eight intermediate hundredths worth only $178, or about $1.80 per hundredth.

The impact of the small 1 percent probability is even more exaggerated when emotions come into play. In an experiment by Rottenstreich and Hsee, a group of students imagined that they could receive either the opportunity to meet and kiss their favorite movie star or $50 in cash. Most preferred cash to a kiss.

But when another group of students was asked to imagine that they could take part in either a lottery offering a 1 percent chance of winning the opportunity to meet and kiss their favorite movie star or a lottery offering a 1 percent chance of winning $50 in cash, most bet on the movie star.

Despite the fact that in certain conditions, the cash was preferred to the kiss, in uncertain situations, the kiss was preferred to the cash.

When making decisions, people are likely to react to the image and affect conjured up by these possibilities, such as the relatively exciting and emotion-laden kiss as compared to the relatively pallid and emotion-free cash, and not to the assessed likelihood of these possibilities.

Setting Emotion Aside

How might emotionally laden consequences influence people's reactions to probabilities? Rottenstreich and Hsee offer the following thought experiment. Picture a terrible car crash involving your closest friends. This image is likely to remain essentially the same-equally vivid and harrowing-whether you are told the chances of such a crash are, say, one in 100, one in 1,000, or one in 100,000.

Moreover, the thought of such an image might make you drive more carefully the next time you get into a car. But in deciding to drive more safely, you are reacting to the mental image conjured up by this scenario-not its stated probability of crashing. Essentially the same process may hold for all emotionally rich possibilities.

In its most extreme form, such a process implies that any probability would be treated like any other: Managers might grossly exaggerate the impact of small probabilities, grossly understate the impact of large probabilities, and be entirely insensitive to any probability variations.

So when an executive is emotionally attached to the possible outcomes, he or she who makes the final decision concerning the product line introduction may inappropriately "ignore" the odds of success and the odds of failure.

"Instead of making a decision that is based solely on the judged likelihood of success, the manager may be swayed by emotional thoughts and affective images concerning how hard the company's team has worked, or how thrilling it would be to read news detailing the success of the project," says Rottenstreich.

A manager who erroneously reacts to these emotions rather than to the judged probability of success, he says, will essentially treat every probability in the same way. "As a result, the manager may be severely overaggressive or severely under aggressive when the probability of success is low," Rottenstreich says.