The Knowledge Depot: Building and Evaluating a Knowledge Management System
Organizations subsist on communication and coordination. Whether the organization is a multi-national conglomerate with a centuries-old history or school children playing basketball at recess, success of the organization depends on the ability of its members to communicate and coordinate. Communication and coordination provide the means to produce and store information that the organization needs. Organizations have goals and are focused on achieving those goals in a way that minimizes the drain on limited resources [March, Simon 1958].
Much of the information generated within an organization is stored within the people in the organization. That is, information about procedures, goals, practices, and history is stored in the people’s memory, desk drawers, and bulletin boards. For decades, this informal method of managing information was, although not ideal, generally adequate. However, recent changes in business, and opportunities in technological advances make such informal methods inadequate.
Traditionally the backbone of an institution is its employees. These workers, who were once a stable force, have changed significantly. Since the knowledge of many institutions resides in its workers, the loss of these employees or the isolation of their knowledge can threaten the core knowledge of the organization, lead to rework and decrease productivity [Garvin 1997]. To compound this problem, we have seen that the rapid changes and increasing complexity of technologies has made it difficult and expensive for all workers to be trained effectively. It is becoming more and more important for workers to share best practices and to form communities that support each other in solving problems.
Additional pressure is also being placed on workers. With the downsizing efforts and mergers that increase the geographic spread of the work locations, many people are working with people over a greater distance. These workers have greater reliance on electronic resources for communication and coordination of their work force than they needed to in the past.
Solutions to address these problems lie in both technological and organizational changes of the institution. This paper briefly summarizes some organizational changes that were made to address these problems. However, the focus the work presented here has been on the technological changes.
A successful solution to these problems is one that facilitates the generation, storage, and retrieval of information among a distributed workforce. The Knowledge Depot is a knowledge management system that facilitates communication and coordination among the over 2000 students, faculty and staff in an innovative training program that involves 25 colleges spread over six states. Students work full-time and attend classes one day a week. Because students work in different places and at different times, face-to-face communication is rare and students use the Knowledge Depot for much of their communication and coordination.
The Next Step Organization and the Knowledge Depot System
The Bell Atlantic Next Step Program is an innovative educational program established to educate Bell Atlantic technicians for advanced technical positions in the corporation. The developers and maintainers of Next Step Program are a collective group of individuals consisting of Bell Atlantic management, Union management, and the faculty of the associated colleges. The curriculum was specifically designed for this program by this group. This program has been deployed at 25 colleges throughout six states in the Northeast and currently has over 2000 members. This organization is a distributed organization spanning colleges throughout New York and New England. Most of the members of this organization are part-time members. Students work full time for Bell Atlantic and go to school one day a week during company time.
Goals and Requirements of the Knowledge Depot System
The authors, who were members of the Advanced Technology group of Bell Atlantic, were chartered to recommend, design and develop knowledge management technologies for this organization. The high-level goals the organization had for such a system were as follows:
These goals fed into the requirements of the system. Requirements were also developed based on the characteristics and limitations of the organization as described below. The design was also strongly influenced by the members of the organization and on the experience of the developers in designing interactive systems. These design considerations are described in a later section.
Many Knowledge Management systems have focused on supporting particular groups or a particular type of knowledge instead of the organization as a whole [see "Related Systems" section below]. Other work has relied heavily on providing content to an entire organization without much attention to the privacy and unique needs of small work groups [Alavi 1997]. Since the Next Step organization consists of many subgroups (e.g. classes of students, faculty of a particular subject), these groups needed individual support. Since the organization as a whole had many common issues, support was also needed at a high level. Therefore the following became a requirement:
Many systems require a large number of knowledge workers to help organize and develop the content. This presents a problem for organizations with limited resources and a long list of other priorities. These organizations need to take advantage of intellectual capital [Putnam 1993] that is developed as part of the work process. One such example would be the communication and coordination of distributed work teams that many employees are now required to do.
Since this organization does not have an extensive budget and thus the system needed could not heavily depend on a large team of dedicated knowledge workers, the following was also required:
Even when the organization as a whole, and individual groups are supported, and content is generated without an unreasonable cost to the organization, a challenge arises. This challenge is in making members of the organization aware all the relevant resources available to them. This awareness mechanism must be based on each individual’s needs. It must also give the individual the freedom to tailor it to his/her preferences (e.g. push versus pull, delivery frequency, and individual interest profile).
Since the Next Step organization is a distributed, dynamic organization, consisting of many part-time members and the frequent addition of new members, it is very important to keep members aware of the online content. This is difficult to perform with a limited staff so the following was also required of the system:
While this paper mainly focuses on technical solutions to the goals and requirements of the organization, there are several interesting organizational changes that took place that aided in the adoption of the Knowledge Depot system. These are mentioned below.
Moving from informal information foraging to communication-supported collaborative learning is an organizational problem, as much or more than it is a technical problem. Providing a technical "solution" to a work group is only as effective as the work group's willingness to use the technology. Moving from an environment in which information exists primarily in individual memories to one in which information is shared requires organizational change as much as technical change. Technical change can support organizational change, but technical change cannot produce a solution without organizational change [Argyris 1993]. Therefore it is important to note here some of the organizational changes that the Next Step Organization made that enabled the technology to be successful.
As outlined in [Boyatzis, Cowen, Kolb 1995], it is necessary to challenge convention and tradition of the status quo in education in order to facilitate change and to make the organization a learning organization. That is just what this group did. In defining what is important to the business, beyond the more traditional curriculum, this group defined a set of “Umbrella Competencies” that were to be woven into each course. These Umbrella Competencies were designed to help students bridge the gap between formal education and real world experience and to focus the students on customer satisfaction. Most importantly they were designed to produce the behavioral, cultural, and organizational changes needed to help create a learning organization. The competencies focus on quality, teamwork, teambuilding, leadership, technology, customer focus, interpersonal skills, and problem solving and instructing skills. The focus on teamwork was very important in the use of our technology. As one student so eloquently stated “A few databases and other physical resources would be practically useless without the sense of camaraderie, teamwork and, yes, friendly competition fostered by the program.”
Recently, many systems have been marketed to support web-based learning environments. While the Knowledge Depot can also be viewed as a system to support web-based learning environments, the focus is significantly different from that of other systems. The Knowledge Depot system focuses on an entire organization whereas other systems developed have focused on supporting a single class.
Some of the systems developed to create web-based learning environments to support educators are TopClass [TopClass Web-based Training], LearningSpace [LearningSpace Lotus Notes Corp.], and WebCT[WebCT, University of British Columbia]. These tools focus on helping a faculty member create online course materials, manage administrative tasks and create an online interactive environment (synchronous and/or asynchronous) for the students and the faculty member. These tools are used to either complement the in-classroom experience or to allow a course to be taught completely online. They are not intended, however, to facilitate sharing of information among instructors or colleges.
The Next Step Organization is unique in that while it spans many colleges, the curriculum is designed to be consistent throughout the organization. Also all the students of the organization have a similar work environment and thus are more likely to share even if they are not at the same college in the same class. Therefore our focus is on sharing materials among colleges so that students and faculty members at a college could benefit from any other college. The result is a stronger focus on sharing, collective repositories and collective discussion areas. Our challenge lay in designing the system carefully to avoid chaos due to the large number of people involved with the program. For example, if each faculty member were able to create their own unique repository for each course they taught in the program, there would be over 220 repositories for each semester, each with a very common thread.
As Davenport and Prusak warn in their book [Davenport, Pruzack 1998], “if you build it, they may not come,” it was important to think about how to design this system so the organization would use it. As designers, we did not have much control over the personal reward systems of the individual users or the management mandate that many [Davenport, Pruzack 1998, Orlikowski 1992]recommend will enhance usage of the technology and therefore we could not motivate our users as such. However, it has been shown that other factors are alsoinvolved with designing and deploying such a system that contribute to its success [Grudin 1994 Orlikowski, Hofman 1997]. A “human-centered” approach [Davenport 1994, Mackay 1990],could strongly affect the usefulness and usability of the system deployed. Consequently, it was important to focus on designing, developing, and deploying such a system. In attempting to do this, it was possible to build on the experience of several researchers in the field of knowledge management systems and groupware who have studied the success and failure of such systems.
Following Grudin’s suggestion [Grudin 1994], it was important to design the system so that there were identifiable benefits to the people who used the system. When an individual uses the system, the benefit gained from this experience shouldencourage them to continue using the system. Therefore the system needed to be designed in such a way that there were benefits to the current users, not just the future users. In doing so, the system will have a better chance of sustaining continued use.
It was also important to support both formal and informal knowledge, making the system flexible enough so that broad content types were supported [Davenport 1994]. In this system there are discussion areas where informal conversations take place. Some of these conversations help surface tacit knowledge by allowing people to describe how to do a task, others are conversations in which members provide support to each other but do not necessarily share knowledge. There are also places where formal documents can be placed such as student projects, faculty course handouts, and reference materials. The problem with such a flexible environment is often with retrieval. As a result search tools are provided that search each area individually as well as a search mechanism that can search multiple databases. Sharing tools were provided that help users stay aware of the content in the various repositories. These sharing tools are described later.
Multiple levels of organization of the content was supported and the system was designed so that knowledge can be structured at any time after it is entered [Shipman, McCall 1994]. It was not desirable to force the content to be too structured but necessary to provide structuring mechanisms so that it could be automatically structured or restructured at a later time. The group repositories (referred to later as Group Depots) were designed to support this flexibility. In other repositories, a minimum structure requirement (e.g. sorted by college/topic) was instilled so that the information and knowledge could be more accessible.
The usage of the system was anticipated to increase by automating as many tasks as possible thus creating a rich set of repositories without user intervention. Member information was generated as much as possible so users would not have to manually enter information that could easily be accessed from other sources. Documents were also automatically collected by catching all email that was sent to a group. This alleviates users from having to remember to store such documents. Finally, users were automatically notified of new content that was placed in the repositories. Users could modify their notification details and schedules but are initially set up with a default notification schedule. These automatic features were installed for all users by default to entice them to use the system [Mackay 1990] and to alleviate users from having to fully understand the system in order to use it.
As Grudin suggested [Grudin 1994], it is best to build upon an already successful application. While this system was being developed, Lotus Notes was being installed as the corporate communication standard at Bell Atlantic. While most of our user populations did not previously use Lotus Notes, this did provide motivation to learn. Building on an application that the user population was learning reduced the overhead of learning to use the system and may even have provided a welcome opportunity for many to practice the Lotus notes skills they were learning elsewhere.
Finally, ideas were borrowed from the field of Participatory Design [Greenbaum, Kyng 1991], Evolutionary Growth [Fischer 1994], the Improvisational model [Orlikowski, Hofman 1997], and the framework specified in [Zimmermann, Selvin 1997]. Much of the system was designed jointly by our team and the managing staff of the Next Step Program. Several student classes were involved in refining the design of the student Group Depot databases. The authors and designers of the system are in contact with program members continually though email and meetings. In addition surveys are conducted at the end of each semester, and data collected on how the system is being used. As a result the organization, its use of the technology, and the technology itself has co-evolved and continue to do so.
The Knowledge Depot System
The requirements were to create repositories to support the entire organization as well as individual work groups, to automate as much information capture and organizing as possible, and to ensure that organization members are aware of the information resources available to them. These requirements were addressed by building four main system components. The requirements and these components are discussed below.
First, the focus was on creating repositories to support the entire organization as a whole as well as to support individual work groups. The Electronic Publication component provides mechanisms to create and maintain general repositories. With this mechanism, members given permission could easily publish content to program members. A group repository component, called Group Depots, was created to support specific groups or a set of groups. This component captures and organizes group communication and other group artifacts. A Communication Network component was created that allows tracking of what groups there are in the organization and how they relate to one another. This makes it easier to keep the group repositories secure and to limit access to electronic publication databases when necessary. This component also provides a way for the users to communicate with relevant groups in the organization and defines the communities of practice for the organization.
Second, since the Next Step organization had limited resources and could not afford a staff of knowledge workers to create and maintain content, it was necessary to take advantage of the intellectual capital that is developed as part of the work process. This was done by automating as much information capturing and organizing as possible. Group mail was collected, stored and organized in the Group Depots. Existing corporate databases were used to populate the databases with member information. Usage log data was collected to publish when users had logged on last, and the communication network data was used to publish users group affiliations. Additionally, the security of Lotus Notes and its authoring capabilities were used to make it easy for users to create and maintain their own online content, thus removing the need for a middle person to collect and publish information.
Third, members of the organization need to be aware of all the relevant resources available to them. This was accomplished by creating the Sharing Mechanisms component. This component kept members aware of the online resources as well as aware of each other. Through the introduction of a subscription feature, individuals can tailor which resources they will be kept aware of and how often.
Below the four main components of the Knowledge Depot system are described in detail. The Knowledge Depot system was built upon the Lotus Notes platform. The Electronic Publication and the Communication Networks mechanisms were developed by using the basic Lotus Notes features of developing databases and creating entries in the Lotus Notes Name and Address book. The Group Depots and the Sharing mechanisms were developed using the Lotus Notes API and C++ agents that provides the ability to expand the basic functionality of Lotus Notes.
In organizations there is an ongoing need to disseminate certain information and knowledge to the entire organization. Traditionally, this content was given to employees in documents and organization-wide mailings. Currently, most organizations have some form of electronic publication for all organization wide materials. It represents large, organization-wide repositories mainly designed for the broadcast of information and knowledge to all members of the organization. The ease of broadcast frequently outweighs the need to get the right information to the right people at the right time. There is a danger of broadcasting information to people who do not need it. We will return to this issue below in discussing communication networks.
Through analyzing what information was currently being sent and whom it was sent to, we created a group of repositories. This is a common electronic information center that consists of a set of Lotus Notes databases, that is accessible to all members of the Next Step organization. The Group Interaction category is a special case. Although a few of the resources in the group interaction are open for all program members to view, this category may be restricted to individual groups. This is discussed later in the Group Depot section.
Several infrastructure elements are important in effective knowledge management technology. One is a common set of software tools [Davenport, Pruzak 1998] for such tasks as email, word processing, and presentations, a second is a hierarchical representation of the groups in the organization. We call this the “Communication Network” that represents a grouping of the individuals who interact with each other in their common activities. These groups of people are often called the Communities of Practice [Lave 1988]. This representation is more complicated than an organization chart especially in a distributed organization where groups of people who are not co-located work together. In the Next Step organization, we needed to figure out what groups of people worked together, what was the group hierarchy if any, what social networks existed that need to be supported, and what networks need to be encouraged. Once an initial representation was made, these groups were then used when creating group repositories, or group depots, as described in the next section.
In creating the group representation, we wanted to avoid two problems. The first is communication isolation and the other is communication overload. Having no groups formally defined for the entire organization could lead to communication isolation. On the other hand, defining groups too broadly could cause people to send out broadcast messages to too many people. If the organization is broken down into clearly defined groups, then mail senders could more easily communicate with the exact audience desired.
For the Next Step organization we followed a set of steps to define the groups. We obtained a list of all the people in the program. We grouped together existing work groups. We then defined other groups whom have a common interest or problem (e.g. all Lotus Notes support people, all people organizing an event, etc.). We then defined a hierarchical group structure from the subgroups. We then defined Lotus Notes groups to represent the group structure and sent out memos to the groups informing them of the structure. Lastly we obtained the commitment of the man aging staff to keep the current group lists up to date.
Group repositories are set up in the Knowledge Depot system to capture group communication and information. We call these the “Group Depots.” Group Depots differ from the other Electronic Publication databases in that they focus on a particular group in the organization and were often kept private to that group. The exact requirements of the Group Depots change depending on a specific group’s needs
The Group Depot can be broken down into two components: a database and a set of agents. The database component is a Lotus Notes database shared by a set of group members. This is a place where group members can read, write and arrange information relevant to the group. Figure 2 is a representation of how agents and users interact with a Group Depot database.
Mail messages that are sent to the entire group are automatically captured and organized in the Group Depot by the Distributor and Categorizer agents. The Distributor agent collects a copy of all messages sent to the Next Step groups or carbon copied to a Next Step group’s depot, calculates which depot the mail belongs to and distributes it to that depot. The Categorizer, continually analyses the messages in each Group Depot database and categorizes mail according to group defined categories called Message Categories. Message Categories have three attributes that define which messages get placed in them. These attributes describe the subject, sender, and recipient fields. Besides being automatically organized, group members can manually organize the messages and modify message categories. This allows the group structure to evolve as the group’s needs evolve. If a user creates a new message category, the Categorizer will search through all existing messages to see if they should be placed into that category. Users can, at any time, manually place messages into a category or remove messages from a category. Messages can appear in more than one category
Initially most Group Depot databases are set up to be only accessible by the members of the group. All group members have equal permissions to create new documents and annotate or recategorize messages.
Awareness and Sharing Mechanisms
In a large organization, it can be very hard to keep track of all the resources available and to keep up with the changes going on in them. Awareness mechanisms are designed to make the program members more aware of the various repositories available to them and to make them more aware of one another. Sharing mechanisms are designed to share the content in the group repositories across several groups or the entire organization.
We have developed awareness mechanisms to make members more aware of online resources as well as each other. To make program members explicitly aware of the online resources, we developed an agent, called the “Welcomer,” that continually checks for new program members. When the Welcomer agent recognizes a new program member, it sends the member a Welcome Memo, that describes all the online resources, that are relevant to the new program member based on their group membership. To help program members know more about each other, we developed a repository that contains basic information about each person in the organization. This includes basic contact information about each person as well as when they were last online and what groups they are a member of. This repository is automatically maintained by our software.
We have tried to satisfy a variety of levels of sharing needs as well as mechanisms for information delivery. We have defined three pull mechanisms. First, when two groups are very closely related and do not have a privacy issue between them, they can share the same Group Depot repository. Second, when groups are interested in each other’s activities and privacy is not a concern among them, we allow them to browse each other’s Group Depot database. This was done with the curriculum faculty groups to enable them to stay aware of changes happening in other curriculums. Third, we enabled the publication of Group Depot information so that a group can make available to people outside the group a set of documents or mail messages created during a particular semester or time period
There are two ways for information to be pushed to a member. If a member is interested in everything about a group, and the group does not object, that person can become a member of that group. By doing so this person would have access to all the group resources and would receive a copy of all the group mail. However, the most interesting push mechanism in this system is what we call the Subscription feature [Kantor, Zimmermann, Redmiles 1997]. This feature allows program members to subscribe to any of the repositories that they have permission to. Upon subscribing to a repository, the users will periodically receive mail messages, that we call subscription messages (see Figure 3), that summarize all the documents in the repository that have been entered or modified. The user can specify the time period in which these messages are to be received and can combine information about several repositories into one message. A subscription message contains summary information about the new or modified documents with a hypertext link to each document. This alleviates the members from having to poll the repositories for new information.
Evaluation of the Knowledge Depot
The Knowledge Depot is a knowledge management system that facilitates communication and coordination within Bell Atlantic's Next Step Program. In order to evaluate such systems, we must use two perspectives: (1) how effective is it as a knowledge management system? and (2) how useful is it to the organization for which it was developed? We address these perspectives below, beginning with the organizational perspective.
The Organizational Perspective
Success on the high-level organizational goals for the Knowledge Depot System is difficult to demonstrate. Clearly, a technological system that does not meet organizational goals cannot be viewed as successful. Equally clearly, the best people to argue that organizational goals were met are the staff, students, and faculty in the Next Step Program. Still, we give our arguments below and are cautiously optimistic that these arguments would be adopted and elaborated on by the Next Step Program members.
The principal organizational goals were: (1) to support and enhance communication among organization members, (2) to document group decisions and history and include sharing mechanisms so that communication can occur across groups and across time, (3) to avoid information overload and information isolation, and (4) to share best practices, avoid rework, and provide an environment that enables teamwork/team building, leadership, and instructing skills.
The first organizational goal was to support and enhance communication among organization members. The usage data we have clearly shows that people use the system.
We collected usage data of the online repositories and individual mail boxes. In addition, surveys are given at the end of each semester to both faculty and students. The return rate for the faculty surveys was 10% for the fall 1996 through the fall 1997 semesters, and was 20% for the spring 1998 semester. The return rate for the student survey was about 20% for the fall 1996 through the fall 1997 semesters, and was 30% for the spring 1998 semester. The program has over 2000 members.
Table 1 shows data from the latest survey asking people how they preferred to do specific types of activities. This survey included both faculty and student members. We sort these activities here into two categories, that which consists of obtaining information about the Next Program itself and that which involved personal and casual correspondence. As would be expected, group depots have little effect on personal and casual communication. However, as we had hoped, they are an equally preferred resource for obtaining program information, which is largely archival.
Table 1. Preference of method to do the listed activities.
To provide further support for the conclusions implied by the survey data, we analyzed usage data. Figure 4 shows the total number of reads/writes/opens to the mailbox databases and the online repositories. In this figure we can see that the usage data matches the survey data in that the Notes mail usage (to groups or individuals) does exceed the usage of the online repositories. This coincides with the fact, found in another survey question, that the activities that they use the online repositories for are done less often than the activities that they use the one-on-one communication for. The fluctuation in usage for mail and the repositories seems to be fairly consistent with each other. These fluctuations match what we would expect for a school year with valleys during the vacation periods. The gradual increase in both represents the increase in the size of the organization each semester.
In summary, the above data indicates that the group depots are used. But, if they are truly used, we would expect use to increase over time. Table 2 shows changes over four survey periods in preferred information source for curriculum information. We focus here on faculty members since faculty were able to browse curriculum group depots for groups they were not members of. Clearly, use of group depots increases substantially, while email use declines as a method of searching for curriculum information.
Table 2. Changes in preferred information source over time for curriculum information.
Furthermore, the fact that they use the Knowledge Depot even though they have other communication options, such as telephone and face-to-face conversations, suggests that participants saw the Knowledge Depot as an effective communication mechanism. However, the strongest argument in support of achieving this goal is the active involvement of many of the students, faculty, and staff in the design and deployment of the Knowledge Depot. Even though these groups may not have communicated much in the past, they saw the Knowledge Depot as their common system to design and to own. The communication and cooperation that may have initiated in system design carried through in system use. The second organizational goal was to find ways to share information across groups and across time. The clearest evidence here is that data that shows that groups do browse the depots of other groups. For example, faculty members browse the curricula of other faculty (see Group Depot usage in Table 2). Students belonging to different labor unions shared information about labor relations. Clearly information was shared across time and across groups. We would also like to claim that the awareness mechanisms, such as the "Welcomer agent" made people aware of information resources. While we believe this to be true, we admit that we have no baseline data that shows the level of awareness prior to introducing the Knowledge Depot system. Collecting such baseline data would be a worthwhile pursuit in future efforts, however with a system in real use, it would not be desirable to leave users unaware of online repositories.
Success of the third organizational goal, avoiding information overload and information isolation, is one that we have to argue for more on the basis of belief than data. Successfully achieving this goal would mean that we defined the communication networks properly. We do know that these groups were used, but we cannot say that they would not have been better used if they were redefined. The staff has responded to requests from the users in group definitions but in future efforts, more work is needed here. One alternative we would like to consider is to allow group and individuals some flexibility in defining group membership dynamically.
Finally, we believe the Knowledge Depot provides an environment in which people can share best practices and an environment that encourages teamwork. There is, for example, evidence that faculty members shared curriculum information and that students formed discussion groups on topics that interested them. However, this is also a place where we lack baseline data and recommend that collecting such data be part of future efforts. As we indicated with respect to the first organizational goal, the strongest argument in support of achieving this goal is the active involvement of many of the students, faculty, and staff in the design, deployment and use of the Knowledge Depot. Their insights into best practices shaped the design and use of the system and the system benefits greatly from their work on the extended design and development team.
The Knowledge Management Perspective
Knowledge management is a common problem but successful knowledge management systems are rare. The need to record and preserve intellectual capital drives organizations to manage knowledge. Such knowledge management, however, relies on individuals to be the repositories and on information foraging to be the primary access method. Systems that build and preserve the intellectual capital of an organization while minimizing individual effort are rare. We have presented one system that meets this goal. Maintaining a focus on organizational goals and on using technology to overcome organizational obstacles are key to successful development and deployment of knowledge management systems.
However, since success is tied closely to organizational goals, how do we know that these goals were defined appropriately? In other words, how do we know that the system is successful for the organization. To examine the system in a more general sense, we turn below to broad discussion of knowledge management.
Davenport and Prusak [Davenport, Pruzak 1998], define what they consider to be primary attributes of successful knowledge management projects. Below we use the attributes to discuss these attributes in reference to our project.
A Final Comment
Evaluating knowledge management systems is a very difficult task. Success can be indicated in many ways. But, the most important indicators come from the people who use the system. As one user indicated, “I would guess that the most obvious way that the Next Step Program has made a difference in the way we perform our jobs has to be in the way we get our information: access to Lotus Notes, the personal networking made possible by having classmates from diverse departments, communication over a vast region through the (group) depots, access to the corporate-wide web, (other corporate repositories), the diverse databases implemented by our fellow students, all add up to a powerful resource only dreamed of before now.”
In this paper, we have described the Knowledge Depot system. This system provided the Bell Atlantic Next Step Program with a knowledge management system with four complementary components. The first, Electronic Publication allows selected members of the Next Step Organization to publish content electronically to its members. The second, Communication Networks, provides a way for the users to communicate with relevant groups in the organization and defines the communities of practice for the organization. The third, Group Depots, captures and organizes group communication and other group artifacts. The last, Sharing Mechanisms, allow users to share group repositories and other online content across the organization.
We argue that collectively these components meet the goals that the Next Step organization set for us. The first organizational goal was to support and enhance communication among organization members. The data we have clearly shows that people prefer to use the system. The second organizational goal was to find ways to share information across groups and across time. Again, there is evidence that shows that groups use the depots of other groups. Success of the third organizational goal, avoiding information overload and information isolation, is one that we have to argue for more on the basis of belief than data. To the extent that we defined the communication networks properly, this goal was met. However, in future versions, we will experiment with allowing group and individuals some flexibility in defining group membership dynamically. Finally, we believe the Knowledge Depot provides an environment in which people can share best practices and an environment that encourages teamwork. There is evidence that faculty members shared curriculum information and that students formed discussion groups on topics that interested them.
Finally, we argue that the Knowledge Depot is a successful knowledge management system. That the Knowledge Depot system has been in use over three years and that over 2000 people choose to use it indicate that this is a very successful knowledge management system.