Background

Chapter 13 - The Game of Internet B2B
Managing IT in Government, Business & Communities
by Gerry Gingrich (ed)
Idea Group Publishing 2003

	Brought to you by Team-Fly

In the present context it would be useful to step back and look at the issues from somewhat different perspectives: social theory and game theory. A broader theoretical perspective allows generalization of these specific findings to the overall process of technology adoption and, indeed, to the development of these technologies themselves.

Fundamentals of Game Theory

Game theory assumes that when players face uncertain outcomes, they seek to maximize their expected payoffs. This "is known as the expected utility hypothesis, and has been controversial since it was first proposed by John von Neumann and Oscar Morgenstern in 1944" (Bierman and Fernandez, 1998, p. 220).

Essentially, any uncertain outcome is modeled as some type of "lottery." The nature of a lottery is determined by the probability of a payoff. A "fair lottery" offers a 50-50 chance of winning or losing, so the expected return over time is zero. In an "unfair" lottery, the chances of losing are greater than the chances of winning (expected returns are negative) and in a "superfair" lottery the chances of winning are greater than the chances of losing. A person who is willing to accept unfair lotteries can be called "risk loving," while a person unwilling to play even a fair lottery is "risk averse." A person's willingness to accept a given type of lottery may change based on the magnitude of the potential payoff or loss (Bierman and Fernandez, 1998, pp. 229 30).

In games of incomplete knowledge, a collection of probability assessments about the state of the world formed from observing the moves of opponents is called the player's "belief profile." Following the expected utility hypothesis, each player's strategy always selects moves that maximize the player's expected utility, given knowledge of the game up to that move, beliefs about the state of the world at that point, and beliefs about the other players' strategies (Bierman and Fernandez, 1998, p. 328). In this class of games there are two roles, which can be called "informed" and "uninformed." "Bayesian updating" of belief profiles during a game consists of simply replacing the prior probability of a state of the world with its posterior probability, after observing the moves of a better-informed player (Bierman and Fernandez, 1998, p. 323).

Beliefs and Roles

Game theory accepts the expected utility hypothesis and reduces roles to the "informed" and the "uninformed" player in order to facilitate the process of constructing mathematical models. Rossiter and Percy (1987) develop a more complete set of roles for purchase decisions. They stress that having an effect on group decision-making requires reaching the individuals who will have the most influential roles in the decision. Their taxonomy of roles (pp. 109 110) is made up of:

• Initiator: gets the overall purchase decision started

• Influencer: uses product information to either promote or retard the overall decision

• Decider: makes the overall "go/no-go" decision

• Purchaser: executes the decision (although there may be scope for change)

• User: ultimate consumer or user of the product

Simon (1959, pp. 51 2) describes a role more generally as:

... a social prescription of some, but not all, of the premises that enter into an individual's choice of behavior. Any particular concrete behavior is the result of a large number of premises, only some of which are described by the role. In addition to role premises there will be premises about the state of the environment based directly on perception, premises representing beliefs and knowledge, and idiosyncratic premises that characterize the personality ... With our present definition of role, we can also speak meaningfully of the role of an entire business firm of decision premises that underlie its basic policies ... The common interest of economics and psychology includes not only the study of individual roles, but also the explanation of organizational roles of these sorts.

The key point here is that considering a company as a player in a game is not simple anthropomorphism. Communities of interest within the organization will determine the action, which is ultimately visible to other players. A company's belief profile is a collective interpretation of the payoffs, probabilities, and actions of the other players.

The Structurational Model of Technology

Orlikowski develops Giddens' theory of structuration into a social model of technology that can be used to add depth to further analysis. Structuration theory focuses on the interplay of "structure" and "action." Structure is a body of social learning embedded in each member of a society (and thus an organization or community as well). "Structure has no existence independent of the knowledge that [human] agents have about their day-to-day activity" (Giddens, 1984, p. 26). That knowledge is gained from monitoring actions and their consequences both intended and unintended. "That is to say, actors not only monitor continuously the flow of their activities and expect others to do the same for their own; they also routinely monitor aspects, social and physical, of the contexts in which they move" (Giddens, 1984, p. 5). "According to the notion of duality of structure, the structural properties of social systems are both medium and outcome of the practices they recursively organise" (Giddens, 1984, p. 25).

The organizational context is defined by three fundamental elements of social interaction: meaning, power, and norms. Meaning is shared knowledge that informs and defines interaction, as distinct from the conventions and rules that govern "appropriate" behavior (norms). Power refers to the asymmetry of resources that participants bring to, and mobilize within, interaction. (Orlikowski, 2001, pp. 64 65) In other words, the rules and resources that constitute structure both enable and constrain action. At the same time, structure is reaffirmed as actors observe and understand what they are doing in the context of producing and reproducing day-to-day social encounters.

Technology is both the product of human action (design) and a medium of action (use). The structurational model of technology relates technology (material artifacts mediating task execution in the workplace), human agents (designers, users, decision makers) and institutional properties of organizations (both internal and environmental) (Orlikowski, 2001, p. 71). It assumes that these elements interact recursively, may be in opposition, and that they may undermine each other's effects (Orlikowski, 2001, p. 75).

Technology constrains and enables action, thus conditioning social practices. Designers embed rules reflecting knowledge about the task being automated and the organizationally sanctioned process for executing a sequence of tasks into a new resource for accomplishing the work. Once created and deployed, technology remains inanimate and, hence, ineffectual unless and until it becomes a medium of human action (Orlikowski, 2001, p. 73). When users conform to a technology's embedded rules they consciously or unconsciously reaffirm the institutional structures in which the technology is deployed. When users do not use the technology as it was intended they may undermine these structures and, hence, the strategic objectives of the sponsors and designers (Orlikowski, 2001, pp. 74 75).

Application

This perspective gives substance to the interplay of the facilitators and inhibitors described above. Organizational norms and shared meaning are the domain of management and strategic vision, which guides the participants in the collective decision-making process. Economies of scale and market expansion are resource issues. The firm's image and competitive position create and constrain strategic vision, as well as resources and alternative uses for them. All of these elements are known to a player in a game, but they are not known to all players, and will be subject to updates based on other player's actions.

While Orlikowski does not consider it explicitly, the structurational model of technology easily incorporates re-engineering designs that are meant to challenge organizational norms and shared meaning. It is important to recognize in this context that the rules to be embedded in a re-designed process are not simply received by the initiators and influencers. The actions they take in defining a set of costs and benefits associated with an option for action may themselves undermine the intent of other rules embedded in the larger process.

Thomas (1994) describes a case where a flexible machining system (FMS) was an investment alternative for a manufacturing company. Upper management had set high standards for return on investment (ROI), but uncertainty about the actual cost of the system and the development team's lack of experience made the exercise "... really silly. We had a number and we hit it" (Thomas, 1994, p. 207). However, " the operations manager gave credence to the R&D group's assessment of the importance of imagery in decision making. He preferred the FMS because 'harder working machines' and fewer people were something he thought his superiors could understand" (p. 203). Indeed, the R&D manager who had supervised the writing of the proposal had made an explicit effort to use the same words and phrases that the operations manager had used in describing an "antiquated factory" run by a competitor (Thomas, 1994, p. 202).

Thomas states, "The net result was that technological choices were made largely on the basis of personal and professional interests and what was perceived to be an archaic social context in the shop that could not be altered without the external pressure of technological change" (1994, p. 210). It is also interesting to note that in this case, the operations manager was promoted to a higher position in the company (Thomas, 1994, p. 211).

The Role of the Expert

Interpretation of action or choice of possible courses of action are frequently based on mediated knowledge rather than first-hand experience. Giddens (1991) points out that while in modern times there are plenty of claimants to authority, in reality there are no determinant authorities but "an indefinite pluralism of expertise" (p. 195).

Specialisation is actually the key to the character of modern abstract systems. The knowledge incorporated into modern forms of expertise is in principle available to everyone, had they but the available resources, time, and energy to acquire it. The fact that to be an expert in one or two small corners of modern knowledge systems is all that anyone can achieve means that abstract systems are opaque to the majority. Their opaque quality - the underlying element in the extension of trust in the context of disembedding mechanisms, comes from the very intensity of specialisation that abstract systems both demand and foster (Giddens, 1991, p. 30).

The domain expert requires a vocabulary for describing the topic area, and a set of rules for combining terms and relations, which allow for extension of the vocabulary (an ontology) (Farhoodi and Fingar, 1997). In structuration terms, a community of experts forms around meaning and norms. When it comes to resources, Williamson's (1975, p. 111) characterization could apply equally to the purveyors of technology, and the technologies themselves:

Reputation, which is to say prior experience, is of special importance in establishing the terms of finance for transactions that involve large, discrete commitments of funds... Faced with incomplete information, suppliers of capital are vulnerable to opportunistic representations. Unable to distinguish between those unknown candidates who have the capacity and the will to execute the project successfully from the opportunistic types who assert that they are similarly qualified, when objectively (omnisciently) they are not, the terms of finance are adjusted adversely against the entire group.

In game theory terms, Williamson casts the resource provider as the uninformed player, and the community of experts as the informed player. In a signaling game, the better-informed player makes a move before the less-informed player; in screening games, the less-informed player moves first. "In short, the screen is essentially a set of hoops set by the uninformed player that the informed player can choose to jump through or not, depending on the rewards offered by the uninformed player" (Bierman and Fernandez, 1998, p. 336). Here, the suppliers of capital set up a screening game by demanding a higher payoff before they will accept the proposition.

Fact vs. Hope

Opportunistic representations are a real danger. They may be internal and deliberate, as illustrated by the FMS case mentioned above where management had set ROI as a screen for acceptable projects. On the other hand, they may be the result of imperfect knowledge on the part of external experts.

For example, when NASA (1998) looked at the acquisition process for Commercial Off-The-Shelf (COTS) software, they "... expected vendor interaction to be simple and to end with the purchase of a product. In reality, ... the [study] team found a strong dependence on bi-directional information flow" (pp. 5-2). The vendor was one more party with whom communication channels had to be established and maintained at different levels throughout the project. Project personnel had to rely on the vendor for a variety of technical issues, but vendor personnel were not always as helpful or available as promised (NASA, 1998, pp. 4 2). This led to reliance on a number of other parties, including other projects that use COTS, independent evaluation teams, and other customers of the vendor (NASA, 1998, pp. 5 2).

In sum, mediated information may add to rather than reduce the risk associated with the adoption decision. Development and deployment of technology requires a team of domain experts, each with its own set of shared meanings. "Narrative coherence with which the reader or viewer can identify becomes the way to commodify prepackaged theorems of how to 'get on' in life ..." (Giddens, 1991, pp. 198 9). We could find an explanation here for the high degree of concern about Internet security, relative to startup and maintenance costs. To information security specialists, these are inseparable: security policies and mechanisms must be established before the system is brought on-line and audited over time to ensure proper implementation and timely evolution as new threats are identified (Northcutt and Novak, 2000, p. 390). However, the horror stories about computer "break-ins" are far easier to relate to than a rigorous routine designed to prevent them. When effectiveness of the measures themselves depends on the expertise of the software and system security vendors relative to the hostile experts attacking the systems, the situation seems even more opaque.

Summary

While dependence on technology does not really increase the risk of a failure that requires unique expertise to fix, adopting new technology requires a change of expertise that negates the value of a particular set of acquired experience. Small wonder then that Wirtz and Wong (1999) found an overwhelming majority of their respondents saw no need to adopt Internet B2B technologies. Multi-organizational systems are doubly abstract, in the sense that they use opaque technological systems to tie together the internal processes of a number of companies. Adoption will ultimately depend on the perception of the game, and the payoffs relative to the risks.

Within the organization, one community of experts that must form is the IT support team. Their expertise will replace the expertise of others that are left behind in the transition. Management guidance is also necessary to re-shape organizational norms. Visible investment of resources must be complemented by visible promotion of new ways to think about the firm and how work gets done. The drive for innovation may come from company norms, or communities of experts within the company.

Market expansion and global reach hint at economies of scale, although another angle on global reach is considered below. In the absence of clear competitive necessity and economies of scale, the management's perception of possibilities to enhance the firm's image will be a key motivation for investment. This will be moderated by the perception of successful implementation of IT applications, both within the firm and externally. Widely publicized concerns about Internet security will increase uncertainty and perceived risk.

	Brought to you by Team-Fly