information that might bear on the outcome in question.^2 For the new product,
this would include consumer surveys, test-market results, the product’s unique
qualities, its price relative to prices of competing products, and so on. The point
is that a subjective probability is not arbitrary or ad hoc; it simply represents the
decision maker’s best assessment, based on current information, of the likeli-
hood of an uncertain event. In this sense, all probabilities—even those based on
frequencies or statistical data—represent the decision maker’s degree of belief.Expected Value
The manager typically begins the process of analyzing a decision under uncer-
tainty by using a probability distribution. A probability distributionis a listing
of the possible outcomes concerning an unknown event and their respective
probabilities. As we saw earlier, assessing relevant probability distributions is
the first step in the manager’s analysis. For example, the manager might envi-
sion the probability distribution shown in the table for the first year’s outcome
of a new-product launch.Outcome First-Year Sales Revenue Probability
Complete success $10,000,000 .1
Promising 7,000,000 .3
Mixed response 3,000,000 .2
Failure 1,000,000 .4This probability distribution provides the best available description of the
uncertainty surrounding the market’s reception of the product. Note the con-
siderable range of outcomes and the high likelihood of failure. (Revenue of $1
million is not enough to justify continuing the product.) Failure is the norm for
even the most promising new products.
From the probability distribution, we can compute the expected value of
the uncertain variable in question. In the preceding example, expected rev-
enue is (.1)($10) (.3)($7) (.2)($3) (.4)($1) $4.1 million.502 Chapter 12 Decision Making under Uncertainty(^2) Any probability forecast is based on the decision maker’s currently available information.
Consequently, if this information changes, so will the probability assessment. Thus, a disap-
pointing market test would lead management to lower its probability assessment of product suc-
cess. The point is that probability assessments are not engraved in stone; rather, they are constantly
being revised in light of new information. In addition, various “experts” often hold different sub-
jective probability assessments about an event based on different information or different inter-
pretations of common information. (In contrast, the objective probability of heads in a single
coin toss is immutable; that is, it is always one-half. Assuming there is no doubt about the fairness
of the coin, this probability will not change with new information, nor will it be subject to dispute.)
We take up the subjects of information acquisition and probability revision in Chapter 13.
c12DecisionMakingunderUncertainty.qxd 9/29/11 1:34 PM Page 502