"I wish these guys would stop fighting and be a little more accommodating of each other's point of view," thought TomHoffmeyer, Vice President of Marketing at the General Mills Company.  Two of his managers, PetePiscopo who held product management responsibilities and Ron Roberts, marketing research manager, were in serious disagreement about whether they should conduct a test market before launching a new brand of RTE cereal.  Ron (the research manager) was recommending a market test before deciding whether to launch the new cereal so as to assess if the main launch would successfully meet its sales, market share, and profitability goals. Pete, always the entrepreneur, felt that the test market could never provide perfect diagnostics and would waste a lot of time and money that they could ill-afford.  From his product manager perspective, he felt that the company would do a lot better going directly to market.  The company had already spent $150,000 in attributed development costs for this new product.

About two weeks ago, Tom had participated in a meeting in which Pete and Ron, both well-respected in the organization, had exchanged some fairly sharp words.  Their disagreement was significant and the discussion grew heated to the point whereRon had insinuated that, as product manager, Pete had become excessively ego involved with the product and was potentially misleading management with over-optimistic market projections. Pete, in turn, had implied that Ron was excessively cautious and was intentionally slowing things down in an effort to drum up internal business for his research organization.Hoffmeyer knew that rumors about this acrimonious meetinghad spread and his supervisors were concerned that the internal dissension was holding upa potentially important addition to the product line.

As he walked back to his office, Tom ran into Chuck Raverty, a marketing analyst who had just joined the team.  Chuck was a newly minted MBA who had already acquired areputation as a smart and helpful team-player.   Tom asked Chuck whether he could help with the situation.  After recruiting Chuck in this facilitator role, Tom had called Ron and Pete and calmed them down sufficiently so restart a conversation.  They had each agreed to write down their respective assumptions to explain how they had reached their different conclusions.  The memos had arriveda couple of days ago and Tom handed them over to Chuck for evaluation. 

In reviewing the rather long-winded and passionate memos, Chuck noted that the two warring managers actually shared many of the same assumptions.  However, they also differed on some other critical assessments of sales and profit potential.  However, the differences appeared to be more a matter of degree versus serious structural differences. Chuck developed a tabulation showing the similarities and differences in the information and opinions.  He was both surprised and relieved to see that Ron and Pete had similar assessments of the test market's likely diagnostics regarding whether the product would reach breakeven sales levels.

Indeed, as Chuck shared this tabulation with Ron and Pete, both were surprised to note these points of agreement.   This alone helped greatly in relieving thetension that had surfaced recently between them.  With the acrimony set aside, they appeared emotionally better prepared to have a constructive conversation.

Chuck's analysisshowed that both managers had thought about the sales and profitability forecasts in terms of a favorable and an unfavorable market scenario.  They had also thought in terms of a most likely sales outcome under each scenario.  However, their assessments of sales and profitability outcomes differed because they had anchored to different benchmarks in making their forecasts.  Ron had provided a more complete analysis of sales and profitability.  Pete's memo did not have a well articulated profitability analysis but showed a high level of product specific market knowledge. Although neither wanted to admit this initially, they agreed that their individual perspectives were probably biased by cognitive, emotional and cultural factors that led them to focus on different benchmarks in developing their forecasts.

As marketing research manager, Ron felt that his past research on consumer receptivity to a new brand of cereal yielded some reliable sales and profitability benchmarks.  Chuck summarized Ron's qualitative commentary as implying that in a favorable market (probability about 70%) expected sales during the three year planning horizon would be between 500,000 and 800,000 units (most likely or mean level 650,000 units).  At this most likely sales level, estimated profit would be $2,650,000.  However, Ron had also estimated that there was a 30% chance that the market would be unfavorable.  If this were the case, the corresponding three year sales figures would be in the range of 300,000 to 500,000 units, with the most likely number being 400,000 units.  The estimated loss in this unfavorable scenario (at the most likely sales level) would be $2,120,000.

Chuck realized that it was the likelihood of this unfavorable downside scenario that was driving the angst that Ron had expressed.  He also noticed some similarities in Pete's thinking as reflected in his assessments.  Wanting to  make sure, Chuckwalked over to Pete's office to see if he could create a framework in which the two sets of assessments were comparably aligned.  The conversation with Pete was quite productive.  First, Pete was comfortable with Ron's framework and carefully assessed his favorable, unfavorable and most likely sales forecasts.  However, he had not though through the relatively likelihoods of the favorable and unfavorable scenarios.  When Chuck probed him about these estimates, the Pete volunteered that he felt that there was an 80% chance that the favorable scenario would obtain.  The likelihood of the unfavorable scenario was 20%.

When asked about the associated sales forecasts, Pete confirmed that in the favorable scenario, he expected sales to lie between 500,000 and 1.1 million units (most likely level 800,000 units).  His sales expectations under the unfavorable scenario were in the 400,000 to 500,000 range.  Chuck pointed out that the sales range that Pete had provided was rather wide under the favorable scenario and relatively narrow under the unfavorable scenario.  After reflecting on this, Pete said that his forecasts were conditioned by the possibility of competitive pricing moves.  If the competitive response was not particularly aggressive, he would expect sales at the upper end of the favorable forecast.  However, an aggressive competitive price cut (even if consumer response were favorable) would produce sales in the lower end of the range.  In any case, he thanked Chuck for pointing this out and said that thinking about it had helped him to be a little more specific about the factors that influenced his forecast.  Pete also thanked Chuck for his help in assessing the profitability levels associated with the most likely sales forecasts in the favorable and unfavorable scenarios.  With a little help from Chuck, he estimated these numbers to be a profit of $4.25 million in the favorable scenario and a loss of $1.1 million in the unfavorable scenario. 

Chuck then compiled the following table to summarize these forecasts and forwarded his report to the two managers.  He made sure that the report contained explanations regarding the assumptions underpinning the forecasts.   Chuck was frankly a little worried that his report cataloging the different opinions may re-ignite the earlier tension.  Rather apprehensively, he suggested a set of times at which the two mangers could meet together with him to talk further about the situation.

Chuck was very pleasantly surprised by the friendly e-mails that he received from both Pete and Ron.  Interestingly, each expressed thanks for helping them to articulate their own assumptions and recognize that the other person had theirown reasons for making their assessments..They did note that they were making different recommendations (Ron to do the test market and Pete to launch without doing the test market).  They had agreed to meet next Monday morning over breakfast.  Each expressed a hope that the meeting would be productive and that they could reach an informed decision about the test market and the product launch. 

Chuck realized that just like his MBA days, he needed to set aside part of the weekend to develop a facilitating framework for the meeting.  As he thought about what he might do, he remembered a course he had taken on managerial decision behavior.  In that course he had worked with a Bayesian framework for determining the value of information to support marketing decision making.  He found his recently filed notes from the course and developed the following course of action.

1.       Construct conditional payoff tables for the marketing research manager and the product manager that would lay out their forecasts for analysis.

2.       Calculate the expected value of perfect information (EVPI) given the respective conditional payoff tables for the marketing research manager and the product manager respectively.

3.       Next, calculate the expected value of imperfect information (where the imperfect information pertains to the 85% accurate market research study).  Chuck noted that it was a blessing that both the marketing research manager and the product manager agreed regarding both the cost and the diagnostic capabilities of the proposed research study.  Otherwise, the analytical problem (not to mention the political problem) would become more complicated.

4.       Compare the expected value of imperfect information (market test) to its cost ($350,000) and then decide whether or not the test was worth doing.

Chuck realized that the steps above would lead him to an expected value based decision rule for risk neutral decision makers.  He recalled from his course that if a risk-neutral analysis led to a decision to do the market test, a risk-averse decision maker would clearly want to do the market test as well.  However, he wondered what he should do if the analysis from a risk neutral decision maker's perspective showed that the test was not warranted.

He suspected that like most managers, both the marketing research manager and the product manager were likely to be risk averse and may view the decision differently from his risk neutral analysis.  However, he also suspected that Ron (as market research manager) and Pete (as product manager) were likely to differ in their degree of risk aversion.  He wondered what he should do to deal with that issue.  However, as he contemplated his Friday evening date, he decided to worry about that decision after he had completed the calculations required in Steps 1-4 above.

Chuck was also thinking aboutsome concerns that Hoffmeyer had expressed about this product launch situation. Hoffmeyer was concerned about the $150,000 in attributed product development costs.  He also believed thatif the new product was not launched, General Mills would have to spend additional market development dollars to support the rest of the product line.  In a favorable market, this could mean an additional $75,000 in media and trade promotions cost.  In a difficult market, the cost could be as much as $125,000.  Chuck wondered if and how these costs should be factored into his analysis.  He decided to get on with the analysis in Steps 1-4 and returnto these issues later.

 Problems

 Consider the General Mills Company case that we discussed in class.

A.  Develop the decision tree representations of the problem as presented in the case.  Show the appropriate probabilities and the payoffs for the product manager and marketing research manager.

B.  Consider the information provided in the final paragraph of the case (p. 4).  Be sure to evaluate which costs matter for the decision at hand.  Develop a revised payoff table and reconstruct the analysis in light of this new information.  What impact does this new information have on your decision?

C.  An innovative market research firm has offered to develop a new market test for about $200,000, which is significantly less expensive than the test that was evaluated earlier.  However, the research firm has asked Chuck Raverty for clear directions regarding whether the new test should focus on lowering the probability of the Type I error, the Type II error, neither, or both.  PP and RR retain their original priors and the profitability estimates.  Given the original EVD, EVPI and EVII calculations, what instructions should Chuck give them?  Explain your answer with relevant calculations.