The Use of Information Markets in Corporate Idea Management

Table of Contents

Table of Contents

List of Figures

Abstract

1 Introduction
1.1 Problem definition
1.2 Objective and progress

2 Theoretical basics and state of research
2.1 Information markets
2.1.1 Prediction markets
2.1.2 Preference markets
2.1.3 Idea markets
2.2 Idea management

3 Methodology

4 Results
4.1 Temporal evaluation
4.2 Applied methodology
4.3 Countries of origin
4.4 Keyword evaluation
4.5 Referred subject areas

5 Discussion

6 Conclusion
6.1 Implications for academia and practitioners
6.2 Limitations and need for further research

References

List of Figures

Figure 01: Overview on information market types

Figure 02: Process scheme of literature selection

Figure 03: Resulting subject areas

Figure 04: Publications by years

Figure 05: Overview Methodology full sample

Figure 06: Origin of empirical data

Figure 07: Researching countries by the time of publication

Figure 08: Publications per country

Figure 09: Countries conducting corporate markets

Figure 10: Number of literature contributions per topic

Figure 11: Literature contributions and related topics

Figure 12: Shareholders as a proportion of the total population (2011)

Figure 13: Classification scheme for information market keywords

Abstract

The use of digital trading platforms tremendously gained on interest during the Corona crisis. The possibility to trade stocks oneself via mobile applications lead to a significant rise in interest in financial markets. However, such trading places are not only used for stock trading. For many years, information markets have been a reliable tool for predicting events, especially in the field of politics, sports or movie business. Due to the high prediction accuracy and the high innovation pressure on companies, the question arises whether such markets are suitable for evaluating ideas in companies. For this purpose, 36 articles from journals and conference proceedings were consulted in the context of a systematic literature review. Thereby, nine topic areas could be found in which reasons against the implementation of information markets for this use case exist. The topics in descending order of the number of assigned articles are: Accuracy, Manipulation, Motivation, Complexity, Costs, Knowledge, Confidentiality, Organization and Legal. The research results were also analyzed according to the geographic profile, the publication numbers over time, as well as according to their keywords. All these points taken together lead to the fact that the adaptation into practice has proceeded only slowly. Furthermore, this study assumes no widespread use of information markets for the use case of idea evaluation in the future.

1 Introduction

During the corona crises the use of online trading platforms tremendously increased. This hype had significant impact on the international financial markets, since it gave access to many, especially young traders, who were not interested in financial markets before (Handelsblatt, 2021). But such online marketplaces cannot only be used for the exchange of shares or other financial products. The possibility of generating a price through the aggregation of supply and demand can be also used in other areas. Such so called information markets could support the decision making in various areas and are already used as forecasting tools. Events where those markets have already been widely applied are political markets, sports markets or movie markets. Other usages contain predictions on terrorism targets conducted by the United States Defense Advanced Research Projects Agency or even predictions on rainfalls in Australia (Horn & Ivens, 2015, p.13).

1.1 Problem definition

The generation and evaluation of new ideas is one of the most important tasks in a company (Soukhoroukova et al., 2012, p.1). While various methods for the phase of idea generation are already well established in industry, the idea evaluation remains a serious challenge for companies (Blohm et al., 2016, p.44). Many of them do not suffer from a lack of ideas, but rather fail in their selection and their implementation, since the process of deciding which ideas are worth pursuing and which are not can be quiet difficult (Soukhoroukova et al., 2012, p.1). As Blohm et al. (2011, p.4) describe, researchers have been examining various methodologies for the selection process and also applied the concept of information markets to the evaluation of new product ideas (Bothos et al., 2008; LaComb et al., 2007; Soukhoroukova et al., 2012), new product concepts (Dahan et al., 2010) and early stage technologies (Chen et al., 2009). Knowledge of future market environments and developments can be helpful in making business decisions, since they could help mitigating the risks of failures and bring a competitive advantage (Horn & Ivens, 2015, p.11).

1.2 Objective and progress

Regarding an industry perspective, information markets have been implemented and tested, especially in large multinational enterprises as Abbott Labs, Arcelor Mittal, Best Buy, Chrysler, Corning, Deutsche Telekom, Electronic Arts, Eli Lilly, Frito Lay, General Electric, Hewlett-Packard, Intel, InterContinental Hotels, Masterfoods, Microsoft, Motorola, Nokia, Pfizer, Qualcomm, Siemens, and TNT (Cowgill et al., 2009, p.1). However this leads to the contradiction, that despite the reporting of a very high forecast accuracy and promising results in general (Bothos et al., 2009, p.3), there has not been a wide establishment of information market approaches for the evaluation of ideas or innovations in companies. This was also reported by Cowgill and Zitzewitz (2015, p.1339) as they pointed out the need for further research by asking: ”Why are corporate prediction markets not more popular, including at firms that have already experimented with them?” Following this question, this study examines:

“Why did information markets for corporate idea evaluation did not spread in industry despite early success stories? “

In order to answer this research question, the literature dealing with the application of corporate information markets for the assessment of ideas, innovations, new product developments or other free non-verifiable decisions is systematically analyzed. The aim is to shed light on the dissemination, the resulting benefits, but also possible downsides of the application of information markets for this purpose within enterprises.

In the following chapter, the theoretical foundations and definitions related to the topics of idea management and information markets will be explained. In particular, the differences between the individual types of information markets are highlighted and briefly explained to create a basis for the subsequent evaluation results. In chapter three, the methodology used in this work is presented, whereby the limiting factors and the search process are explained in detail. Subsequently, the results of the systematic literature review are presented and the answers to the research question posed are discussed. Finally, a conclusion is given which contains academic as well as managerial implications and points out the limitations of this work as well as ideas for further research.

2 Theoretical basics and state of research

This chapter explains the concepts that are fundamental to understanding this work. As Tziralis and Tatsiopoulos (2007, p.76) report, there exists no unique and globally adopted definition of the concept and mechanism and therefore the terminology used to address this concept is rather wide. This work is based on the definitions given by Kamp and Koen (2012, p.40) and therefore distinguishes between prediction, preference and idea markets as subsets of the term information market. Figure 01 provides an overview of the different fields of application and the corresponding accuracies of these three types.

2.1 Information markets

The terms information or virtual stock markets describe a whole set of digital marketplaces, which aggregate information through the formation of prices (Bothos et al., 2009, p.2). Regardless of all differences of the individual types of such marketplaces, they are all based on two fundamental principles.

First, the efficient market hypothesis, which Hayek (1945, p.530) considers the most efficient way of aggregating dispersed information from various market participants. Thus, in an efficient market, the price that arises from supply and demand reflects all available information (Bothos et al., 2009, p.2; Horn & Ivens, 2015, p.17). Second, the wisdom of the crowd. The concept of wisdom of the crowd argues, “that aggregation of individual information across a group, even if biased, leads to more accurate predictions and better decisions than those made by any single or small set of individuals” (Chen et al., 2009, p.50). Therefore multiple authors suggest using methods, which rely on wisdom of the crowd for the idea evaluation process or as decision support system for companies in general (Blohm et al., 2011, p.2; Bothos et al., 2009, p.10). According to Kamp and Koen (2012, p.41), in contrast to traditional trading places, information markets do not trade physical resources, but various types of information. “These types of information can be predictions of certain future events, estimates of consumer preferences, or idea rankings” (Kamp & Koen, 2012, p.42).

The trading on a digital marketplace should offer a game-like process, through which the participants are more motivated, compared to conventional evaluation methods (Bothos et al., 2009, p.12). The differences between the individual types of information markets can be traced back to the events, whereby the terminology is not defined uniformly (Tziralis & Tatsiopoulos, 2007, p.76). As a result Ahlstrom-Vij (2019, p.32) determines the distinction between traditional information markets and so-called self-resolving markets, whereby the substantive distinction is not different from that made by Slamka et al. (2012, p.470), who talk about first and second generation prediction markets or Jones et al. (2009, p.291), who distinguishes between verifiable and unverifiable markets.

Abbildung in dieser Leseprobe nicht enthalten

Figure 01: Overview on information market types

Source: Kamp and Koen (2012, p.41)

2.1.1 Prediction markets

The most common type of all information markets (Jones et al., 2009, p.291; Kamp & Koen, 2012, p.42) became popular with the Iowa Electronic Markets, where markets for predicting the results of elections, movie successes, stock prices and more are running (Berg & Rietz, 2003, p.79). Most famous have been the markets regarding political elections, where many studies showed the superiority over other forecasting methods (Horn & Ivens, 2015, p.12). This high prediction accuracy was also shown in studies, where prediction markets were used as a forecasting tool in business decisions, for example in predicting sales numbers (C. Plott & K. Chen, p.17). Other application areas in which prediction markets are widely used, besides political or business contexts, are sports markets or movie markets (Horn & Ivens, 2015, p.13; Soukhoroukova et al., 2012, p.2).

A decisive difference to preference or idea markets is the coupling of the outcome for the individual trader to the occurrence or non-occurrence of the real event associated with the prediction market (Horn & Ivens, 2015, p.17; Kamp & Koen, 2012, p.42). Users buy contracts, which are bound to certain events in the future. If the event occurs the user receives a certain payoff and otherwise there is no payoff (Blohm et al., 2016, p.29). A big advantage of prediction markets is the high accuracy with only a comparatively small number of participants (Blohm et al., 2011, p.6). Kamp and Koen (2012, p.41), for example, estimate the accuracy of prediction markets at 0.94-0.99, based on a review of published papers. They also report that prediction markets conducted by companies also have a very good accuracy (Koen, 2014, p.90).

2.1.2 Preference markets

The first type of the second generation prediction markets are preference markets (Slamka et al., 2012, p.470). They aim at determining the preferences of a group of people and therefore are used to decide which features in a new product are preferred by a potential customer group (Kamp & Koen, 2012, p.42). To do so, concepts of these potential products are traded via the online marketplace with the result of a single price for every concept. These prices can then be seen as filter for identifying the winning product concept (Dahan et al., 2011, p.498).

The main difference to prediction markets is, that the outcome of the information market cannot be validated with the result in the real world, since not every concept will be launched as final product. Therefore the validation, which also decides about the incentives, can only be done via the comparison with other methods like surveys or expert panels (Kamp & Koen, 2012, p.43). According to Kamp and Koen (2012, p.89), preference markets which measure consumer preferences have an overall accuracy of 0.70 to 0.85.

2.1.3 Idea markets

Idea markets are also a subset of information markets without observable outcome (Kamp & Koen, 2012, p.41; Slamka et al., 2012, p.470). They target at evaluating the success of different alternatives. In many definitions mentioned by Dahan et al. (2011, p.470), LaComb et al. (2007, p.246) or Soukhoroukova et al. (2012, p.2) idea markets are not only designed for evaluating a fixed set of ideas or concepts, furthermore own ideas can be introduced. Therefore, these markets combine creation and evaluation of ideas in one single instrument. The evaluation also relies on the same trading concepts as mentioned in prediction and preference markets, since the underlying principles of market efficiency and the wisdom of the crowd are still the same (Soukhoroukova et al., 2012, p.2).

As in preference markets, information about the real success of the ideas will not be available in the short term, or even be never available for the ideas, which were filtered out (McDonagh & Buckley, 2014, p.78). Therefore, several different possibilities for determining the outcome have been developed, such as expert committees, using the volume weighted average prices or, in some cases, even waiting for several years to validate with the real life result (Soukhoroukova et al., 2012, p.6). The consensus between idea markets and these different kinds of evaluation methods is described as moderate by many authors like Soukhoroukova et al. (2012, p.10) or Kamp and Koen (2012, p.10). The latter report an accuracy for idea markets of 0.10 to 0.46, which shows the high fluctuation in the consensus with the compared methods.

2.2 Idea management

The final aspect, which is explained is the concept of idea management, since this study focuses on the usage of information markets for the evaluation of various kind of new ideas. According to Brem and Voigt (2007, p.306) “idea management can be seen as a subprocess of innovation management with the goals of effective and efficient idea generation, evaluation and selection.” The importance of innovations for a company can be seen in the definition of van de Ven (1989, p.20), who refers to innovation as “the process of bringing any problem solving idea into use […] it is the generation, acceptance, and implementation of new ideas, processes, products or services”.

However idea management is not only about creativity or generating new ideas, even if the ideation phase is quite important (Sandstrom & Bjork, 2010, p.2). It also has a focus on providing the right structure, frames and tools for efficient working (Aagaard, 2013, p.449). This goes along with Brem and Voigt (2007, p.306) who see the key issue in the “structured collection and generation of both internal and external ideas; as well as the logical evaluation and selection of those that offer the biggest potential for future corporate success.”

Even though several authors state that there is no generally accepted definition for idea management (Aagaard, 2013, p.449; Brem & Voigt, 2007, p.306), the related aspects and tasks show the importance of idea management as a whole and also of a structured evaluation and selection process. A key challenge at this is the balance between having a structured approach or framework, while not hindering the required flexibility to conduct projects (Verganti, 1999, p.372). This is also supported by Christensen et al. (2008, p.98) who point out several challenges for structured innovation processes, in particular for more radical and risky ideas.

3 Methodology

As methodology for this study a systematic literature review is chosen, since it enables a way of evaluating and summarizing existing literature with minimal bias, high efficiency and consistency (Mulrow, 1994, p. 597-598). The methodology used in this work follows the ruleset from Tranfield et al. (2003, p.209) to exclude bias and to be able to guarantee transparency, completeness and reproducibility of a systematic literature analysis. Therefore, the selection of literature is limited by several factors, which are explained in the following (Fink, 2014, p.50-51). Furthermore, the methodology of this work is based on the six-step process as used, for example, in Kiel et al. (2016, p.679), Rashman et al. (2009, p.466) and Soni and Kodali (2011, p.241). This procedure is described below to ensure clear audit trails and to justify all stages of the process as demanded by Rojon et al. (2021, p.12).

Step One: Selection of the period under consideration:

For the literature selection, it was determined that only literature published between January 1988 and December 2020 would be considered. The year 1988 was chosen as the beginning of the period under consideration, because in this year the IOWA Electronic Markets started operation, which represents the first widespread prediction market and is seen by many authors as the cornerstone for further developments in the field of information markets (Gruca et al., 2003, p.96). The upper limit of December 2020 was chosen to guarantee that no changes in the sample were done during the evaluation.

Step Two: Selection of databases:

Scopus, Business Source Complete (EBSCO), ABI/Inform, Econbiz, Emerald insight and Web of Science were determined as relevant databases. In addition to the hits contained in these databases, the snowball method was applied in the further course. Thereby the bibliographies of relevant literature are searched for further hits. This procedure has already been used in literature analyses such as: David and Han (2004), Kiel et al. (2016), Soni and Kodali (2011) and Webster and Watson (2002) and is considered accepted. Fink (2014, p.28) even recommends using the bibliographies to search for further literature in order to be able to cover all topics.

Step Three: Selection of publication types:

The basic assumption is that articles published in highly cited and ranked journals are of higher quality, and thus the quality of the systematic literature review increases when focusing on these journals exclusively (Kiel et al., 2016, S. 679). This understanding of quality is based primarily on the rigorous review and correspondingly large selection process to which the articles are subjected to before the publication in one of these journals. The research area of the application of information markets for decision-making processes, which are not directly verifiable, only occurs to a very limited extent in highly ranked journals. It was therefore decided to not to limit the selection of publications on the basis of a required ranking, but to also include conference proceedings and working papers.

Step Four: Selection of articles

In order to achieve an accurate and complete picture of the literature dealing with information markets for idea management, the search string used in this work was created on the basis of several previously published works dealing with related topics. The resulting code that was used for the queries in the mentioned databases can be divided into three central components:

Block 1: (idea* OR innovation* OR invention* OR develop* OR suggestion*)

The first block results from the application area considered in this thesis. For the selection of suitable terms, which also establish the connection to the idea management, the works of Gerlach and Brem (2017, p.145), Lasrado et al. (2015, p.182), Jensen (2012, p.2) and Mikelsone and Liela (2015, p. 108) were used as sources for the keywords.

Block 2: (apprais* OR assess* OR estimat* OR judg* OR investigat* OR analy* OR rat* OR select* OR ident* OR screen* OR filter*)

The second block refers to the activity to be performed by the information market, the evaluation of ideas. For this, some of the most common synonyms were considered.

Block 3: ("prediction market" OR "idea market" OR "preference market*" OR "virtual market*" OR "virtual futur* market*" OR "information market*" OR "virtual stock market*" OR "idea future*" OR "decision market*")

The third block stems from the problem of inconsistent nomenclature in science, as described in chapter two. Therefore, works were considered that deal with the structure of research in the field of information market research. In particular, the works of Horn, Ivens et al. (2014, p.96), Klingert (2017, p.51-52) and Tziralis and Tatsiopoulos (2007, p.76-77) were used for the keyword search.

Compounded, this results in the following code:

TITLE-ABS-KEY (( idea* OR innovation* OR invention* OR develop* OR suggestion* ) AND ( apprais* OR assess* OR estimat* OR judg* OR investigat* OR analy* OR rat* OR select* OR ident* OR screen* OR filter* ) AND ( "prediction market" OR "idea market" OR "preference market*" OR "virtual market*" OR "virtual futur* market*" OR "information market*" OR "virtual stock market*" OR "idea future*" OR "decision market*" ))

It should be noted that the code was related to the title, the abstract and the keywords. This increases the number of hits, since the queries are thus performed independently of each other. The used boolean links are chosen wide enough to sort out only definitely not relevant literature (Fink, 2014, S. 25). The search string was written in English, since the dominant publication language in science and also in this subject area is English.

Overall, the queries based on the code, as well as the selected parameters for the time period, publication types, and language restriction resulted in a hit count of 1312. Based on the reading of the titles, the number of relevant literatures was reduced to 176 hits. After subtracting all duplicates as well as reading the abstract of each remaining hit, 22 final results remained. In the context of this evaluation stage, it was also examined whether the information market acting in the respective work, is a preference or idea market or a classical prediction market. Therefore, publications dealing exclusively with classical prediction markets, such as politics, sports results or sales forecasts for a short time horizon were removed. The restrictions on literature published in English, the type of publication and the time period were also retained in the snowball method, which contributed further 14 hits. In the end 36 hits serve as the basis for this work (see Figure 02).

Abbildung in dieser Leseprobe nicht enthalten

Figure 02: Process scheme of literature selection

Source: own illustration

Subsequently, all 36 results were read in detail and the topics contained were extracted. This resulted in 70 different subtopics, which were assigned to nine subject areas. The terms of the sub-topics and subject areas were, as defined by Locke et al. (2020, p.12) described on the basis of the data created and in the course of the evaluation of the data specified and refined as the resulting framework became more clarified.

The resulting areas (see Figure 03) are Knowledge, Confidentiality, Legal, Organization, Manipulation, Motivation, Accuracy, Complexity and Costs. When creating these topics, special attention was paid on ensuring that they were formulated as clearly as possible.

[...]