Decision Tree Analysis for National Environmental, Inc.
National Environmental, Inc. estimates that its new product, ESA-C, has a 40% chance of being a major winner, a 35% chance of a medium winner, and a 25% chance of being a loser. However, before the company makes a production commitment, it would like to further investigate the situation, because these initial estimates [prior probabilities] may not be accurate. They are considering seeking the advice of a market researcher to conduct a special survey.
From previous experience, it is known that such a special survey can predict success, predict failure, or be inconclusive. Statistically, it is known that of all the new products that were successful [major winner], 65 percent of the time the surveys correctly predicted success, 25 percent of the time the surveys were inconclusive, and 10 percent of the time the surveys predicted failure. For those products that were medium winners, the surveys predicted success 35 percent of the time, 45 percent of the time the surveys were inconclusive, and 20 percent of the time the surveys predicted failure. For those products that were losers, the surveys predicted success 10 percent of the time, 40 percent of the time the surveys were inconclusive, and the surveys predicted failure 50 percent of the time.
The marketing department estimates that if ESA-C is a major winner, the payoff is expected to be $800,000. If the product is a medium winner, the payoff is expected to be $400,000. However, if the product is a loser, the company will incur a loss of $750,000.
The company expects to pay the market researcher $35,000.
The company needs to make several decisions.
(a) Draw the decision tree. Label all of the branches and include appropriate probabilities. Be sure to complete the tree by indicating outcomes [payoffs, costs] at the end of the final branches.
(b) Show the revised probabilities tables.
(c) Identify and show the value of the each node. What is your optimal strategy recommendation? Explain your decision.
(d) Determine expected value of sample information and explain what it means in terms of this situation.
(e) Determine expected value of perfect information based upon information prior to receiving survey outcomes.