Case $[$Dumond I$]$

$"$So that's the simplified version of the decision tree based on what appears to be the critical issues,"

Nancy Milnor concluded. "Calculating expected values, it looks as though we should introduce the

new product. Now, I know that we don't all agree on the numbers in the tree, so why don't we play

around with them a little bit. I've got the data in the computer here. I can make any changes you

want and see what effect they have on the expected value."

Nancy had just completed her presentation to the board of directors of DuMond International,

which manufactured agricultural fertilizers and pesticides. The decision the board faced was

whether to go ahead with a new pesticide product to replace an old one or whether to continue to

rely on the current product, which had been around for years and was a good seller. The problem

with the current product was that evidence was beginning to surface which showed that the

chemical's use could create substantial health risks, and there even was some talk of banning the

product. The new product still required more development, and the question was whether all of the

development issues could be resolved in time to meet the scheduled introduction date. And once the

product was introduced, there was always the question of how well it would be received. The

decision tree that Nancy had presented to the board captured these concerns.

The boardroom was beginning to get warm. Nancy sat back and relaxed as she listened to the

comments.

"Well, I'll start," said John Dilts. "I don't have much trouble with the numbers in the top half of

the tree. But you have the chance of banning the current product pinned at $30$ percent. That's high.

Personally, I don't think there's more than a $10$ percent chance of an out$-$and$-$out ban."

"Yeah, and even if there were, the current product ought to be worth $$300,000$ at least," added

Pete Lillovich. "With a smaller chance of a ban and a higher value, surely we're better off with the

old product!"

"Well, I don't know about you two," said Marla Jenkins. "I think we have a pretty good handle

on what's going on with the current product. But I$\text{'}$d like to play the new product a little more

conservatively I know that the values at the ends of the branches on the top half of the tree are

accounting's best guesses based on a complete analysis, but maybe they should all be reduced by

$$100,000$ just to play it safe. And maybe we should just set the probability of high sales equal to $50$

percent regardless of the delay."

Steven Kellogg had been involved in the preliminary development of the new product more than

anyone else. He piped up$,$ "And the delay is actually more likely than no delay. I$\text{'}$d just reverse

those probabilities so that there's a $60$ percent chance of a delay. But I wouldn't make any changes

on the lower part of the tree. I agree with Marla that we have a good idea about the performance of

the current product and the prospects for a ban."

"I don't think it matters," countered Lillovich. "The changes John and I suggest make the current

product look better than it does in Nancy's analysis. Marla's and Steven's changes make the new

product look worse. Either way, the effect is the same."

Nancy had kept track of the comments and suggested changes. She sat down at the computer and

started to enter the changes. After a few moments, she grinned and turned to the board. $"$In spite of

your changes," she said, "I believe I can persuade you that DuMond should go with the new

product."

Case $[$Dumond II$]$

Nancy Milnor had returned to her office, still concerned about the decision. Yes, she had persuaded

the directors that their disagreements did not affect her analysis; her analysis still showed the new

product to be the appropriate choice. The members of the board, however, had not been entirely

satisfied. The major complaint was that there was still too much uncertainty. Could she find out

more about the likelihood of a ban, or could she get a better assessment from engineering regarding

the delay? What about a more accurate sales forecast for the new product?

Nancy gazed at her decision tree. Yes, she could address each of those questions, but where

should she start?

$[$Hint$]$ Compute the EVPI of $(1)$ information about delay; $(2)$ information about ban; $(3)$

information about sales; $(4)$ information about delay & ban; $(5)$ information about delay & sales; $(6)$

information about sales & ban; and $(7)$ information about delay, sales, & ban. What information is

uninformative given which information?