Research commentary: introducing a third dimension in information systems design--the case for incentive alignment.

Publication: Information Systems Research

Publication Date: 01-SEP-01

Bookmark this article Print this article Link to this article Email this article Digg It! Add to del.icio.us RSS

COPYRIGHT 2001 Institute for Operations Research and the Management Sciences

Prior research has generated considerable knowledge on information systems design from software engineering and user-acceptance perspectives. As organizational processes are increasingly embedded within information systems, one of the key considerations of many business processes--organizational incentives--should become an important dimension of any information systems design and evaluation, which we categorize as the third dimension: incentive alignment. Incentive issues have become important in many IS areas, including distributed decision support systems (DSS), knowledge management, and e-business supply chain coordination. In this paper we outline why incentives are important in each of these areas and specify requirements for designing incentive-aligned information systems. We identify and define important unresolved problems along the incentive-alignment dimension of information systems and present a research agenda to address them.

(Information Systems Design; Incentive Alignment; Distributed Decision Support Systems; Knowledge Management; Supply Chain Coordination)

1. Introduction

Organizations in the new digital economy face unprecedented challenges with the rapidity of change in both the competitive environment and the technology. To operate effectively, real-time decision making requires inputs from multiple participants with differing knowledge, skills, and objectives. To better leverage an organization's intellectual asset for value generation, knowledge management has become the center of many companies' core competency. And increasingly, the shift towards e-business demands that organizations restructure their business models and processes so that information and products can move smoothly across the value chain. Critical to the success of all these business activities and initiatives are information systems designed to meet the challenges of today's business.

Given the increasingly important role of information systems (IS) in organizations, IS-design issues have been the focus of a substantial amount of IS literature. Previous research dealt with at least two dimensions in the design and implementation of information systems to support organizational processes. The software engineering dimension addresses the challenges of creating a cost-effective implementation of a system that is reliable, is easy to modify when users need change, and can be upgraded to new hardware platforms. The user-acceptance dimension is based on the evaluation by system users in terms of its relevance, usefulness, ease of use, satisfaction with the outcomes, the ability to exchange information with other participants, etc.

However, a system that takes into account only principles from the above two dimensions will not necessarily lead to a successful organizational outcome. A personal anecdote illustrates this phenomenon. As part of its knowledge-management initiative, a big consulting firm created a new knowledge repository where consultants could record their knowledge gained from past experience, so that the knowledge could be used by their associates facing similar situations. The organizational objective was to leverage the consultants' collective knowledge for value generation. In marketing this knowledge repository, a consultant from this company boasted that by hiring his firm, a client was hiring not just one consultant, but rather a network consisting of thousands of consultants who had made their knowledge available in this knowledge repository. When asked whether he had contributed his knowledge to the system, the consultant answered that he would have liked to, but that he had not yet done so because he had been too busy. So, while the knowledge-repository software may have qualified as a well-built system from a software engineering perspective, and the users may have perceived the system to be useful, the system did not lead to a satisfactory organizational outcome because there was no incentive for consultants to make the effort to input their individual knowledge into the system's knowledge repository.

We believe that as organizational processes are increasingly embedded in information systems, one of the key considerations of many business processes--organizational incentives--should become the third dimension of any information systems design and evaluation. We call this third dimension incentive alignment. It addresses the high-level design issues that recognize the interests and incentives of the users participating in the process (e.g., users' own objectives may differ from the corporate objectives), the differences in the distribution of information across the users, and the desirability of the eventual collective choice from the organization's standpoint. There are two fundamental, interrelated issues in this dimension. First, given the organizational process embedded in the information system, are organizational incentives in place for users to use the system as intended? That is, will the users make the effort to produce correct information? Will they input the true and accurate information they have? Or can they gain from distorting information? Second, will the use of the information system for a business process yield outcomes that contribute to reaching the organization's objectives? Is it robust against information misrepresentation from its users? Systems can be fault tolerant and easy to use but may fail to contribute to the organizational goal. A design methodology is needed that ensures that the new system and the embedded organizational process will be aligned with the organizational goals. Consequently, an information system should be evaluated along this dimension as well.

Incentive issues have become important in many IS areas, among which are distributed decision support systems (DSS), knowledge management, and e-business supply chain coordination. In this paper we outline why incentives are important in each of the three areas and specify requirements for the design of incentive-aligned information systems. We argue that the collection of processes that are eventually implemented need to satisfy the requirements of how it affects the alignment of user incentives to organizational objectives.

The rest of the paper is organized as follows. Section 2 provides a background on how the three dimensions of information systems design relate to each other, and how the third dimension has been treated in the literature. Sections 3, 4, and 5 focus on the specific incentive requirements in distributed DSS, knowledge management, and supply chain coordination, respectively. Section 6 provides a research agenda with research questions that need to be addressed in designing incentive-aligned information systems. The last section concludes the paper.

2. Background

The research on IS design and evaluation is voluminous. For the purpose of illustrating our position, we classify the attributes that are currently used to evaluate information systems into two dimensions--although such classification is by nature an oversimplification.

The first dimension used to evaluate and design information systems has its roots in the 1960's (Naur and Randell 1969, Buxton and Randell 1970), and the literature dealing with this dimension is huge (for example, Souza 1990, Banker and Kauffman 1991, Henson and Hughes 1991, Swanson et al. 1991, Subramanian and Zarnich 1996, Dekleva and Drehmer 1997). We call it the "software engineering" dimension, because most issues considered along this dimension are concerned with the features of the software product itself (Pressman 2000). The quality of a product is measured with respect to (1) whether the software functions correctly under every possible contingency, (2) whether the system can be transported to other platforms, (3) whether the program code has been well documented in case the system needs to be updated in the future, (4) whether the standards of modularity and architectural design have been adhered to, and (5) whether the whole development effort has been carried out while controlling the overall cost. In terms of analyzing the system's ergonomic design, user involvement was considered as well, but the primary focus of this first dimension was the design of the software itself.

By contrast, the attributes of the second dimension take the system user as the focal point. The Technology Acceptance Model (TAM) (Davis et al. 1989) and the models of cognitive fit (Vessey and Galletta 1991) and task/technology fit (Benbasat et al. 1986, Goodhue 1995) are examples of attributes by which systems are evaluated in the second dimension. The user is central in these theories, while the models investigate and analyze what system characteristics explain the success or failure of a new technology adoption from the user perspective.

Attributes that we classify as the third-dimension deal with the effect of incentives and rewards on the system's outcome. For example, in the TAM model, frequency of use would be a measure of technology adoption. In the third dimension, frequency of use would be complemented by how its use contributes to the overall organizational goal: If the system is not well designed, frequency of use might well lead to inferior outcomes at the organizational level, even if the system is perceived as being useful by an individual user. More specifically, third-dimension attributes concern incentive alignment, i.e., when the system has embedded features that induce its users to employ the system in a manner consistent with the design objective, and hence the organization's overall goals such as profitability and value generation. In game-theoretic terms we call a system incentive aligned when a user's dominant strategy and the preferred user behavior correspond from an organizational perspective. That is, the agent can still freely determine his own behavior and use of the...

Read the full article for free courtesy of your local library.