Homework Help is Here – Start Your Trial Now!
Business
Marketing
Mention at least 5 specific points from this article

Mention at least 5 specific points from this article

Principles Of Marketing
Principles Of Marketing
17th Edition
ISBN:9780134492513
Author:Kotler, Philip, Armstrong, Gary (gary M.)
Publisher:Kotler, Philip, Armstrong, Gary (gary M.)
Chapter1: Marketing: Creating Customer Value And Engagement
Section: Chapter Questions
Problem 1.1DQ
See similar textbooks
icon
Related questions
Question

Mention at least 5 specific points from this article 

New Product Blockbusters: THE MAGIC AND SCIENCE OF PREDICTION MARKETS Teck-Hua Ho Kay-Yut Chen N ew products are engines of growth for many firms. Successful new products are often sources of long-term competitive advan- tage. Indeed, some firms' survival depends on their ability to manage new product launches. In 2004, the average percent of firm's total sales attributable to new products developed within the last three years is about 33%. An extensive longitudinal study suggests that new product introductions increase a firm's long-term financial performance and market value. Thus, the importance of new product development cannot be over- emphasized. A significant portion of a firm's financial resources and managerial attention are spent to develop extensions to existing products and to create new products. More and more products are launched every year.' As much as 8% of a firm's sales are spent in the undertaking of these endeavors. A survey by IEEE Spectrum shows that the top 100 R&D spenders, taken as a whole, spent $254 billion in 2004. In fact, Forrester Research reports that more than 90% of CEOS feel that new product innovation is very or extremely important to growth. The same research study indicates that more than 50% of senior executives are dis- satisfied with the returns on new product innovation. There are significant dif- ferences across firms in new product performance. The PDMA study finds that the best performers in new product development achieve twice-as-high sales than the worse performers. Hence, it is not only possible, but also critical to maximize the return of new products. Two powerful ways of improving the return of new products are to invest in only the most promising new product ideas and to improve supply planning before products launch. Most firms fail to correctly pick new product winners- only one out of every five new product launches is successful." Also, firms fre- quently are unable to capitalize on the successes of a new product blockbuster New Product Blockbusters: The Magic and Science of Prediction Markets because of poor demand forecasts. For example, Nintendo launched the new console Wii in November 2006 with huge product shortages and many frus- trated fans. Similarly. Apple Computer had to push back their international launch date of iPod Mini because of supply constraints. Hence, it is important to discover and manage new product blockbusters. There are two common approaches to predicting new product demand before a launchi. The first is to survey target customers about their purchase intentions. Products that have high purchase intention are selected and launched. This approach has been used widely in the screening of new product ideas. It has two problems. First, surveys do not motivate customers to reveal their true purchase intentions. As a consequence, the collected data can be quite noisy. Hence, the link between stated purchase intention and ultimate purchase behavior is weak, and the associated demand forecast is inaccurate. Second, most people are imitators and rely on others to learn about the potential benefits. of new products. Hence, when surveyed, and without learning from early adopters, they give biased response of their purchase intentions. The second approach to predicting new product demand involves the pooling of experts' opinions. It is used widely in the fashion industry where new product demand is highly uncertain. Under this approach, a group of experts are asked to state their opinions and the average opinion is used to gauge the suc- cess of the new product. This approach has three problems. First, the pool of available experts is usually small. Hence there is considerable variance in the forecast. Second, opinions are typically weighted equally independent of expert's knowledge. Ideally, individuals with better knowledge should be assigned more weight. Third, experts' opinions may not be independent of each other because they may rely on same information source. In this article, we describe a novel approach to screening new product ideas and predicting their demand. Under this approach, individuals are moti- vated financially to participate in an organized market with well-defined rules. The goal of a prediction market is to aggregate relevant information from multi- ple and diverse people. After the new product is launched, the market rewards. participants based on their forecast accuracy. The prediction market addresses the potential problems of the survey- based and expert-based approaches. First, participants are compensated for accu- racy in forecasting. Second, everyone can learn from others about the potential of a new product idea through the markets. Such learning allows individuals to update. their beliefs and develop a better forecast. Third, the price discovery process naturally weighs accurate information more heavily. The same price discovery process also removes redundant and dependent information Teck Ho is the William Halford Jr. Family Professor of Marketing at the Haas School of Business at UC Berkeley. Kay-Yut Chen is a principal scientist at Hewlett- Packard Laboratories. sources appropriately. Fourth, prediction markets can accommodate many par- ticipants at a minimal incremental cost. Once the system is built, it can be used. on a continuous basis. ▶
Transcribed Image Text:New Product Blockbusters: THE MAGIC AND SCIENCE OF PREDICTION MARKETS Teck-Hua Ho Kay-Yut Chen N ew products are engines of growth for many firms. Successful new products are often sources of long-term competitive advan- tage. Indeed, some firms' survival depends on their ability to manage new product launches. In 2004, the average percent of firm's total sales attributable to new products developed within the last three years is about 33%. An extensive longitudinal study suggests that new product introductions increase a firm's long-term financial performance and market value. Thus, the importance of new product development cannot be over- emphasized. A significant portion of a firm's financial resources and managerial attention are spent to develop extensions to existing products and to create new products. More and more products are launched every year.' As much as 8% of a firm's sales are spent in the undertaking of these endeavors. A survey by IEEE Spectrum shows that the top 100 R&D spenders, taken as a whole, spent $254 billion in 2004. In fact, Forrester Research reports that more than 90% of CEOS feel that new product innovation is very or extremely important to growth. The same research study indicates that more than 50% of senior executives are dis- satisfied with the returns on new product innovation. There are significant dif- ferences across firms in new product performance. The PDMA study finds that the best performers in new product development achieve twice-as-high sales than the worse performers. Hence, it is not only possible, but also critical to maximize the return of new products. Two powerful ways of improving the return of new products are to invest in only the most promising new product ideas and to improve supply planning before products launch. Most firms fail to correctly pick new product winners- only one out of every five new product launches is successful." Also, firms fre- quently are unable to capitalize on the successes of a new product blockbuster New Product Blockbusters: The Magic and Science of Prediction Markets because of poor demand forecasts. For example, Nintendo launched the new console Wii in November 2006 with huge product shortages and many frus- trated fans. Similarly. Apple Computer had to push back their international launch date of iPod Mini because of supply constraints. Hence, it is important to discover and manage new product blockbusters. There are two common approaches to predicting new product demand before a launchi. The first is to survey target customers about their purchase intentions. Products that have high purchase intention are selected and launched. This approach has been used widely in the screening of new product ideas. It has two problems. First, surveys do not motivate customers to reveal their true purchase intentions. As a consequence, the collected data can be quite noisy. Hence, the link between stated purchase intention and ultimate purchase behavior is weak, and the associated demand forecast is inaccurate. Second, most people are imitators and rely on others to learn about the potential benefits. of new products. Hence, when surveyed, and without learning from early adopters, they give biased response of their purchase intentions. The second approach to predicting new product demand involves the pooling of experts' opinions. It is used widely in the fashion industry where new product demand is highly uncertain. Under this approach, a group of experts are asked to state their opinions and the average opinion is used to gauge the suc- cess of the new product. This approach has three problems. First, the pool of available experts is usually small. Hence there is considerable variance in the forecast. Second, opinions are typically weighted equally independent of expert's knowledge. Ideally, individuals with better knowledge should be assigned more weight. Third, experts' opinions may not be independent of each other because they may rely on same information source. In this article, we describe a novel approach to screening new product ideas and predicting their demand. Under this approach, individuals are moti- vated financially to participate in an organized market with well-defined rules. The goal of a prediction market is to aggregate relevant information from multi- ple and diverse people. After the new product is launched, the market rewards. participants based on their forecast accuracy. The prediction market addresses the potential problems of the survey- based and expert-based approaches. First, participants are compensated for accu- racy in forecasting. Second, everyone can learn from others about the potential of a new product idea through the markets. Such learning allows individuals to update. their beliefs and develop a better forecast. Third, the price discovery process naturally weighs accurate information more heavily. The same price discovery process also removes redundant and dependent information Teck Ho is the William Halford Jr. Family Professor of Marketing at the Haas School of Business at UC Berkeley. Kay-Yut Chen is a principal scientist at Hewlett- Packard Laboratories. sources appropriately. Fourth, prediction markets can accommodate many par- ticipants at a minimal incremental cost. Once the system is built, it can be used. on a continuous basis. ▶
The Basics of Prediction Markets Intellectual History The idea of soliciting inputs from diverse individuals to improve decision making dates back to the dawn of civilization. The Lord of Menchang, in the Period of Warring States in China (around 300 B.C.), housed three thousand guests in order to tap into advice and expertise from a diverse group. Over two thousand years later, in the fall of 1906, at the annual West of English Fat Stock and Poultry Exhibition, 800 people entered into a contest to guess the weight of a fat ox. The group consisted of a few experts and many laymen. To the surprise of everyone, the average guess (1197) was phenomenally close to the actual weight (1198). Many other similar and amazing examples were documented showing that large groups of individuals consistently outperform experts. These examples share two common traits. First, the number of participants is large. Second, participants come from diverse backgrounds and have independent sources of information. In the modern world, companies have diverse employees and hence pos- sess the promise of tapping into this power. However, this potential is seldom realized. The most common way of gathering input is to conduct a meeting. This method is plagued by several problems. First, members in the meeting may not have incentives to provide unbiased information. Worse yet, they often have incentives to provide biased input. Second, members often yield to their superi- ors because of a hierarchical power structure. Third, there is no systematic way to assign relative importance to each input. As a result, whoever argues most eloquently usually has his or her input weighted significantly more. However, a person's ability to communicate may not have any direct bearing on whether they have relevant information. Economists have long wrestled with this information aggregation prob- lem. In 1948, Edward Chamberlin conducted the first economic experiment to determine whether the market can aggregate demand and supply information. Subjects were provided monetary incentives to buy and sell a fictitious item. Half the subjects were sellers who had different costs of production, and the remaining were buyers who had different values for the item. Sellers were paid in real money based on profit (price x quantity sold cost) and buyers were com- pensated based on net surplus (values - price x quantity bought). The values and costs were designed to reflect both a linear demand and supply function. No one subject, however, was aware of the entire demand and supply functions and hence the predicted market price. The subjects were free to negotiate on a one- to-one basis in a decentralized fashion. If this decentralized market were able to aggregate demand and supply information, the price would be at the intersec- tion of the aggregated demand and supply functions. Despite monetary incen- tives, the market failed to aggregate information and yielded the predicted price. Vernon Smith was a subject in Chamberlin's experiment. Smith recog. nized that market rules can have a dramatic effect on a market's ability to aggre- gate information. In Chamberlin's experiment, the decentralized nature of the 146 UNIVERSITY OF CALIFORNIA, BERKELEY VOL 50, NOI FALL 2007 market did not allow subjects to learn from each other. Consequently, in his 1962 experiment, Smith retained most of Chamberlin's design except for the use of a different trading rule." In Smith's design, offers made by buyers and sellers as well as the transaction prices were posted publicly. This change, while simple conceptually, made a powerful impact. Subjects could now learn from the mar. ket and adjust their behavior accordingly. As a result, this centralized market successfully aggregated demand and supply information to yield the predicted. price. This seminal work laid the scientific foundation for subsequent research on the design of markets. In 2002, Smith won the Nobel Prize in economics for the study of alternative market mechanisms using laboratory experiments. Chamberlin's and Smith's experiments illustrate several important market design principles. First, incentives drive subject behavior. Subjects trade to make money but their trades reveal information to the market. Second, a common metric (i.e., transaction price) is necessary to capture this market information. The most updated information is always reflected in the current market price. Third, a market should be transparent to encourage learning. This is typically. done by making all market activities public. Building on these principles, Charles Plott and Shyam Sunder created the first prototype of the modern prediction market in 1982 and further enhanced it in 1988. Instead of using the market to aggregate demand and supply informa- tion, their market was designed to disseminate and aggregate individual diverse information about the value of a stock. The market prices captured that aggre- gate information and could be used to forecast the value of a stock. The basic design is as follows. Subjects were asked to bet on three stocks. Only one of the three would pay cash at the end of the experimental session. Subjects were pro- vided with partial information about the winner. For example, some might be told that the winner was one of the two stocks (hence the remaining one stock could never be the winner). Subjects were also given some initial allocation of these stocks and seed money. They were then asked to trade these stocks in a market. Subjects could post public offers to buy and sell any stock in real time. The prices of the three stocks captured all the aggregate information of the sub- jects. The market worked well and was able to predict the winning stock reliably. This finding established the potential of the innovative use of markets to predict future events. Consequently, their design became the gold standard for future prediction markets. The first field application that received widespread recognition is the lowa Electronic Market (IEM), which is still running today. The IEM is not-for-profit and operated by the University of lowa Tippie College of Business. The IEM has been focusing on predicting the outcomes of political events, such as presidential elections. Its participants were drawn from the general public and were allowed. to use their own money in the market. However, a speculator can only bet up to $500 in the IEM. The following is an example of how it works. The 2008 U.S Presidential Election prediction market has three outcomes: Democratic Party nominee, Republican Party nominee, and an independent candidate. Spectators can bet on who will become the president. The winning stock pays one dollar CALIFORNIA MANAGEMENT REVIEW VOL SQ NOI FALL 2007 147
Transcribed Image Text:The Basics of Prediction Markets Intellectual History The idea of soliciting inputs from diverse individuals to improve decision making dates back to the dawn of civilization. The Lord of Menchang, in the Period of Warring States in China (around 300 B.C.), housed three thousand guests in order to tap into advice and expertise from a diverse group. Over two thousand years later, in the fall of 1906, at the annual West of English Fat Stock and Poultry Exhibition, 800 people entered into a contest to guess the weight of a fat ox. The group consisted of a few experts and many laymen. To the surprise of everyone, the average guess (1197) was phenomenally close to the actual weight (1198). Many other similar and amazing examples were documented showing that large groups of individuals consistently outperform experts. These examples share two common traits. First, the number of participants is large. Second, participants come from diverse backgrounds and have independent sources of information. In the modern world, companies have diverse employees and hence pos- sess the promise of tapping into this power. However, this potential is seldom realized. The most common way of gathering input is to conduct a meeting. This method is plagued by several problems. First, members in the meeting may not have incentives to provide unbiased information. Worse yet, they often have incentives to provide biased input. Second, members often yield to their superi- ors because of a hierarchical power structure. Third, there is no systematic way to assign relative importance to each input. As a result, whoever argues most eloquently usually has his or her input weighted significantly more. However, a person's ability to communicate may not have any direct bearing on whether they have relevant information. Economists have long wrestled with this information aggregation prob- lem. In 1948, Edward Chamberlin conducted the first economic experiment to determine whether the market can aggregate demand and supply information. Subjects were provided monetary incentives to buy and sell a fictitious item. Half the subjects were sellers who had different costs of production, and the remaining were buyers who had different values for the item. Sellers were paid in real money based on profit (price x quantity sold cost) and buyers were com- pensated based on net surplus (values - price x quantity bought). The values and costs were designed to reflect both a linear demand and supply function. No one subject, however, was aware of the entire demand and supply functions and hence the predicted market price. The subjects were free to negotiate on a one- to-one basis in a decentralized fashion. If this decentralized market were able to aggregate demand and supply information, the price would be at the intersec- tion of the aggregated demand and supply functions. Despite monetary incen- tives, the market failed to aggregate information and yielded the predicted price. Vernon Smith was a subject in Chamberlin's experiment. Smith recog. nized that market rules can have a dramatic effect on a market's ability to aggre- gate information. In Chamberlin's experiment, the decentralized nature of the 146 UNIVERSITY OF CALIFORNIA, BERKELEY VOL 50, NOI FALL 2007 market did not allow subjects to learn from each other. Consequently, in his 1962 experiment, Smith retained most of Chamberlin's design except for the use of a different trading rule." In Smith's design, offers made by buyers and sellers as well as the transaction prices were posted publicly. This change, while simple conceptually, made a powerful impact. Subjects could now learn from the mar. ket and adjust their behavior accordingly. As a result, this centralized market successfully aggregated demand and supply information to yield the predicted. price. This seminal work laid the scientific foundation for subsequent research on the design of markets. In 2002, Smith won the Nobel Prize in economics for the study of alternative market mechanisms using laboratory experiments. Chamberlin's and Smith's experiments illustrate several important market design principles. First, incentives drive subject behavior. Subjects trade to make money but their trades reveal information to the market. Second, a common metric (i.e., transaction price) is necessary to capture this market information. The most updated information is always reflected in the current market price. Third, a market should be transparent to encourage learning. This is typically. done by making all market activities public. Building on these principles, Charles Plott and Shyam Sunder created the first prototype of the modern prediction market in 1982 and further enhanced it in 1988. Instead of using the market to aggregate demand and supply informa- tion, their market was designed to disseminate and aggregate individual diverse information about the value of a stock. The market prices captured that aggre- gate information and could be used to forecast the value of a stock. The basic design is as follows. Subjects were asked to bet on three stocks. Only one of the three would pay cash at the end of the experimental session. Subjects were pro- vided with partial information about the winner. For example, some might be told that the winner was one of the two stocks (hence the remaining one stock could never be the winner). Subjects were also given some initial allocation of these stocks and seed money. They were then asked to trade these stocks in a market. Subjects could post public offers to buy and sell any stock in real time. The prices of the three stocks captured all the aggregate information of the sub- jects. The market worked well and was able to predict the winning stock reliably. This finding established the potential of the innovative use of markets to predict future events. Consequently, their design became the gold standard for future prediction markets. The first field application that received widespread recognition is the lowa Electronic Market (IEM), which is still running today. The IEM is not-for-profit and operated by the University of lowa Tippie College of Business. The IEM has been focusing on predicting the outcomes of political events, such as presidential elections. Its participants were drawn from the general public and were allowed. to use their own money in the market. However, a speculator can only bet up to $500 in the IEM. The following is an example of how it works. The 2008 U.S Presidential Election prediction market has three outcomes: Democratic Party nominee, Republican Party nominee, and an independent candidate. Spectators can bet on who will become the president. The winning stock pays one dollar CALIFORNIA MANAGEMENT REVIEW VOL SQ NOI FALL 2007 147
Expert Solution
steps

Step by step

Solved in 3 steps

SEE SOLUTIONCheck out a sample Q&A here
Blurred answer
Recommended textbooks for you
Principles Of Marketing
Principles Of Marketing
Marketing
ISBN:
9780134492513
Author:
Kotler, Philip, Armstrong, Gary (gary M.)
Publisher:
Pearson Higher Education,
Marketing
Marketing
Marketing
ISBN:
9781259924040
Author:
Roger A. Kerin, Steven W. Hartley
Publisher:
McGraw-Hill Education
Foundations of Business (MindTap Course List)
Foundations of Business (MindTap Course List)
Marketing
ISBN:
9781337386920
Author:
William M. Pride, Robert J. Hughes, Jack R. Kapoor
Publisher:
Cengage Learning
Marketing: An Introduction (13th Edition)
Marketing: An Introduction (13th Edition)
Marketing
ISBN:
9780134149530
Author:
Gary Armstrong, Philip Kotler
Publisher:
PEARSON
MKTG 12:STUDENT ED.-TEXT
MKTG 12:STUDENT ED.-TEXT
Marketing
ISBN:
9781337407595
Author:
Lamb
Publisher:
Cengage
Contemporary Marketing
Contemporary Marketing
Marketing
ISBN:
9780357033777
Author:
Louis E. Boone, David L. Kurtz
Publisher:
Cengage Learning
SEE MORE TEXTBOOKS