A Content-Incentive-Usability Framework for Corporate Portal Design

INTRODUCTION

The knowledge based theory of the firm argues that firms obtain competitive advantage by creating, storing and applying knowledge (Jayatilaka, Schwarz, & Hirschheim, 2003). According to Grant and Baden-Fuller (1995), a firm’s ability to leverage knowledge held by members in the organization is dependent on first, the ability of the firm to create an infrastructure to access this knowledge, transfer it and make it available to others. A second determinant is the extent to which the knowledge that is captured matches with the product domain of the firm.

Enterprise Information Portals have emerged as gateways to streamline information access in firms (Kim, Chaudhury, & Rao, 2002). The first service they provide is access to transactions with the various information sources scattered across the enterprise, such as structured databases, e-mail servers and document repositories. A second service is access to data and knowledge from both internal and external information sources, such as the World Wide Web (WWW). Finally, these portals allow users to interact with other users to perform activities that require team collaborations.

The discussion above indicates that a knowledge portal (KP) is a significant component of an enterprise information portal, and can contribute to a firm’s competitive advantage. In this work, we present a multidimensional framework we term the content-incentive-usability (CIU) framework for KPs to analyze the challenges in building and utilizing KPs.

THE CIU FRAMEWORK The Content Dimension for KPs

The content dimension deals with the determination of the content that should be presented on the KP (what should be presented) and the process of creation of the content (what are the challenges facing this content creation?). We subdivide this dimension into the following subdimensions: elicitation and translation of tacit and explicit knowledge, the integration of structured and unstructured data and the creation of a knowledge ontology to enhance availability.

Elicitation and Translation of Tacit and Explicit Knowledge

According to Nonaka and Takeuchi (1995), tacit knowledge embodies beliefs and values, and is actionable. In contrast, explicit knowledge is codifiable into artifacts such as documents, or multimedia formats. Both are essential for organizational effectiveness.

The transmission of knowledge from one individual to another can take the forms shown in Table 1.

Of the possibilities shown in Table 1, the elicitation of tacit knowledge from experts, and the codification into explicit knowledge represents an important task in the creation of a KP. Eraut (2000) found that elicitation task was easier if:

Table 1. Conversion of knowledge

• there was a mediating object that experts were used to, such as a drawing, a picture or a graph

• a precedent of regular mutual consultation existed between novices and experts

• a training or mentoring relationship was part of the cultural and behavioral expectations in the organization

• informal meetings were held, where “riskier” comments could be made

• there was a perceived potential crisis or change

The degree to which a KP allows the translation of knowledge will influence the final quality of content. Table 1 lists some example technologies that can be used to facilitate the conversions. For example, if we need to capture the tacit knowledge of an expert into a KP, we need to make this tacit knowledge explicit, which can be facilitated by conversations with the expert. The explicit knowledge may then need to become tacit within other users in order to transfer the expertise, and this process can be enhanced if the explicit knowledge is presented on the KP in a form that is easy to visualize.

Integration of Structured and Unstructured Data

Every organization has a large amount of data scattered in sources such as structured databases, e-mail, documents, blogs and newsgroups set up for specific user groups. A major challenge in constructing a KP is the integration of this information. The use of semistructured data to integrate heterogeneous data sources has been shown in several works (Fernandez, Florescu, Levy, & Suciu, 2000; Garcia-Molina et al., 1995). We characterize the issues that need to be addressed in this integration at different layers: the physical layer, the syntax layer and the semantic layer. This is similar to the approach used in Jin, Decker, and Wiederhold (2001) which uses integration, semantic, composition and generational layers.

The physical layer involves the composition of the files that store this data. These files include relational database management system (DBMS) files, word processed documents in various formats and text based or HyperText Markup Language (HTML) files for e-mail, blogs and newsgroups. Part of the challenge is that in most cases, these “islands of information” are not touched, and an automated integration mechanism needs to be created for real-time updating of the KP from these multiple feeds.

The syntax layer deals with the representation of the same information in different formats. For example, information on the same customer may be scattered and/or duplicated across multiple relational DBMSs, documents, blogs newsgroups and e-mails. Duplicated information may have different labels, so that one system may use the customer id as the unique identifier, while another may use the customerac-countnumber for the same purpose. The usage of extensible Markup Language (XML) (Glavinic, 2002) has greatly simplified the mechanism of automation. However, firms still face the organizational challenge of creating a common XML schema that can be fed from these multiple streams. Examples of existing XML schemas that may be used include the TSIMMIS approach in Garcia-Molina et al. (1995) for structured data and the resource description framework (RDF) (Jin et al., 2001) for semistructured information.

The semantic layer deals with the inference of meaning from the data. We propose that one way to accomplish this is to link the data to processes performed by the end-user of the KP. A second method to accomplish this is to create metacategories of the data that map to a knowledge ontology. For example, information on customers, purchases, products and promotions may be combined into a “selling assistant” screen that can be part of the KP. In order to create meta-categories, the meaning of the data needs to be understood. The semantic layer feeds into the creation of a knowledge ontology, described next.

The Knowledge Ontology in a KP

The question of what defines knowledge needs to be answered if knowledge is to be codified and made available. Examples of knowledge include reports and charts from structured data, summary statistics on unstructured data (such as the number of e-mails sent to a customer), and data mining into templates (which are part of the ontology) from blogs, newsgroups and documents. The aim here is to match the knowledge ontology to the product domain and the organizational structure of the firm, to increase efficacy of the KP (Marwick, 2001). For example, in a process driven organization, the knowledge ontology may stem from process descriptions that are already developed. In a functional organization, in contrast, the knowledge ontology would be better off incorporating the functional areas such as sales, marketing, accounting and operations.

Many ways to develop ontologies have been suggested. Some suggestions include using text classifiers (Woods, Poteet, Kao, & Quach, 2006), allowing individual employees to add to an existing list of terms (Amidon & Macnamara, 2003), and forming expert subgroups of employees to develop key words to be incorporated into the ontology (Markus, 2001). However, using these methods individually to develop ontologies can create problems. In the case of text classifiers, this method only allows for ontologies that use existing documents. It is important to share other forms of knowledge such as lessons learned (Gaines, 2003; Gill, 2001; Hanley & Malafsky, 2003; Holsapple & Jones, 2004). This type of knowledge may not be represented in a documented format at the time the ontology is created and key terms may be missed.

A potential problem of allowing individual employees to simply add to an existing list is the organization may end up with so many “key” terms that nothing can be grouped. For example, if one employee uses the term “business re-engineering” and another employee uses the term “organizational redesign” and each added their own term to the list of organizational terms, then the knowledge categorized as “business reengineering” and the knowledge categorized as “organizational redesign” may not be grouped together.

Forming expert subgroups to develop an ontology may solve the above problem. However, now there is the problem of novices not knowing enough to search for the correct key word (Markus, 2001). If the employees are unable to utilize the system designed to do this, then only those who already possessed the knowledge would use the system.

The Incentive Dimension for KPs

Historically, companies have driven their employees to excel through competition (Van Alstyne, 2005). This practice has resulted in employees hoarding their knowledge in order to keep a competitive edge over their coworkers. In this new era of knowledge management (KM), there has been an organizational shift to knowledge sharing. In order for organizations to fully utilize and benefit from the knowledge within the organization, they must find ways in which to encourage employees to share their knowledge (King, 2006). In addition, organizations need to provide means for which the employees can easily participate in knowledge sharing. These activities of securing knowledge sharing efforts and structuring knowledge sharing efforts encompass the knowledge coordination class of activities (for further information on this KM class of activities, see Holsapple & Jones, 2005).

Obtaining management understanding and buy-in of knowledge sharing is clearly needed before any efforts to motivate other employees will be successful (Dorfman, 2001; Lai & Chu, 2002; Lapre & Van Wassenhove, 2001; Massey, Montoya-Weiss & O’Driscoll, 2002; Mullich, 2001). It is important for managers to understand the goal and potential results of sharing knowledge (Delio, 1998). If top management does not buy-in to the idea, they will have difficulty “selling” it to their employees. A lack of enthusiasm from top management can send a confusing signal to the employees. This type of confusion can even lead to employees banding together to deliberately not comply with the knowledge sharing philosophy (Dorfman, 2001). One way to obtain management buy-in is to institute a pilot study of the knowledge sharing program (Massey et al., 2002; Mullich, 2001; O’Dell, 2000). Displaying the success of a pilot group to managers will exemplify the potential benefits to their own areas. This will also provide the managers with support, and perhaps even passion, when trying to motivate their employees to participate.

Social exchange theory indicates that individuals will only contribute when there is an expectation of some future benefit. According to this theory, organizations will need to find ways to illustrate to employees the potential returns of sharing their knowledge (Markus, 2001). Therefore, practices such as rewarding employees for sharing their knowledge with others (Bose, 2002; Liebowitz & Chen, 2003), describing just how that knowledge sharing can be of benefit at both the individual and the organizational level (Delio, 1998; Department of Navy, 2001), publicly recognizing “team players” (Delio, 1998; O’Dell, Elliott & Hubert, 2003), and rewarding employees for participating in a knowledge community (Smith & McKeen, 2003) are ways which companies may motivate employees to participate.

The Usability Dimension for KPs

A large body of literature exists on evaluating and enhancing the usability of computer systems in general (Nielsen, 1993; Shneiderman, 1998). Typical constructs include the learnability of the system (how long does it take to reach a steady state of proficiency?), the efficacy (error rates made by users when performing benchmark tasks), the efficiency (how quickly can users perform benchmark tasks) and the subjective satisfaction of the user. While the first four are clearly measurable, subjective satisfaction can be measured in several different ways. It has been investigated in terms of attitude towards use in many studies (Chou, Hsu, Yeh, & Ho, 2005; Heijden, 2003). Usefulness and ease of use are deeply rooted in attitude towards use. Perceived usefulness is the degree to which users believe that a Web portal will enhance their performance, and perceived ease of use is the degree to which users believe a Web portal will be free of effort (Chou et al., 2005). Usefulness and ease of use have been found to have a significant impact on a users’ intention to use a Web site (Heijden, 2003; Lin, Wu & Tsai, 2005).

User acceptance has also been investigated in terms of data quality and knowledge distribution (Chou et al., 2005). Data quality means the information provided by the Web portal must fit the use of the consumers and generate useful information for the users’ decision-making. Knowledge distribution deals with the need for users’ to use industry Web portals to facilitate employees’ growth and cross-department knowledge sharing. Heijden (2003) also found perceived enjoyment to influence user acceptance of Web portals. Enjoyment is the extent to which using a Web portal is perceived to be enjoyable on its own.

As an example of usability evaluation in the area of Web portals, Yang et al. (2005) developed and validated an instrument to measure perceived subjective service quality of Web portals. The instrument focused on five key dimensions of service quality: (1) usability, (2) usefulness of content, (3) adequacy of information, (4) accessibility, and (5) interaction. Service quality can be seen as a dimension of user acceptance. The five measures of service quality can therefore have an impact on acceptance of Web portals. Usability is related to user friendliness, and it is primarily identified in terms of layout, Web site structure, user interface, appearance and visual design, clarity, and ease of navigation. Usefulness of content is the value, reliability, accuracy, and currency of the information provided by the Web portal, whereas adequacy of information is completeness of the information provided by the Web portal. Accessibility of the Web portal involves availability and responsiveness of the Web site. Finally, interaction exists between the users and service providers’ employees, and users and the Web site, and among peer users of similar products.

Figure 1. Dimensions of the content-incentive-usability framework

FUTURE TRENDS AND CONCLUSION

The main dimensions and subdimensions of the CIU framework are summarized in Figure 1.

The CIU framework can be utilized in several ways. From a practical perspective, it serves as a checklist for organizations who are exploring the implementation of a KP. The discussion of each of the subdimensions in this work should provide prescriptive guidance on increasing the impact of the KP on the performance of the firm. Thus, focusing only on the content without providing incentive or making the portal usable may reduce the chances of success. A three- pronged approach that addresses all three dimensions will increase the potential impact of the portal.

From a theoretical standpoint, the CIU framework serves to provide perspective in the different areas of research related to KPs. Thus, future research projects can be more easily put into perspective with other work, by utilizing this framework to align the project with a particular dimension and subdimension.

KEY TERMS

Incentive: A tangible reward provided to perform a task, such as knowledge sharing.

Learnability: The degree of ease with which a system (such as a knowledge portal) can be learned so the user reaches an acceptable state of proficiency.

Structured Data: Data that follows a predefined format and is stored in a database, such as a relational database.

Unstructured Data: Data that is stored in the form of free-text or images, without a predefined format to help in its access.

Usability: The degree to which an artifact (such as a knowledge portal) is easy to use and adds value to the user.

User Efficacy: The degree to which a user can perform a benchmark set of tasks on a system (such as a knowledge portal) without error.

User Efficiency: The amount of resources (such as time) required by a user to perform a benchmark set of tasks on a system (such as a knowledge portal)