A Case Study: GBS Practitioner Portal
The GBS Practitioner Portal is shaping how IBM's Global Business Services (GBS) continues to evolve with their knowledge management needs. GBS makes up nearly half of IBM's global work force, with more than 150,000 practitioners worldwide. Delivering the right tools, templates, and practical information is a monumental task. The GBS Practitioner Portal team, in partnership with several other teams within GBS, is looking to leverage the latest in Web 2.0 and Social Networking technologies to try to address that challenge.
KM is nothing new to the GBS organization. An organization of this size has to be good at creating, capturing, and sharing information in order to streamline processes and obtain the repeatability required for effective delivery. KnowledgeView is GBS's worldwide knowledge-sharing solution and contains business solutions, engagement experiences, proposals, marketing materials, deliverables, practice aids, and thought leadership materials.
Even with a well-defined repository of documents and information, the Learning and Knowledge Management (L&K) team knew that they had to do more. Other repositories and libraries of information existed that needed to be tapped, searches needed to be federated across these other sources of information as well as within the primary source of KnowledgeView, and information needed to be tagged as to relevance and value it might bring to another user. From this was born KnowledgeView Lite, or the GBS Practitioner Portal, which has several simple but wide-reaching goals. The primary goal of the portal is to improve the ability to find relevant, quality content and experts by
- Providing a simplified user interface.
- Utilizing a Google-like search functionality that includes content from all relevant IBM sources.
- Improving the culture of participation and leveraging the power of the organization via the integration of key social computing and expertise location tools.
- Proactively pushing content via available RSS feeds.
- Highlighting key practice content in business-driven portlets and providing a palette of available portlets to customize the user experience.
Beyond that, however, there are even higher goals to increase the search effectiveness by leveraging user feedback and applying social tagging, ratings, and bookmarks to items. This allows the portal to
- Promote quality content.
- Highlight key content within search results.
- Apply user interaction data to the content lifecycle to promote top quality and automated archiving.
GBS understands that the individual practitioner is one of IBM's greatest assets. The knowledge and collaboration capability that can be provided by each individual is essential in moving the art and science of service delivery forward. The GBS Practitioner Portal is a major step in making achieving that goal (see Figure 1.11).
Figure 1.11 GBS Practitioner Portal
How do you deliver targeted content to 150,000 service practitioners? No portal can be everything to everyone. No single portal instance, at least. GBS has more than 100 service lines ranging across multiple technologies and industries. Not every service line can be profiled as a top-level page and still maintain some sense of importance within the overall structure. The obvious approach is to attempt some form of personalization to target information to specific users based on job role, project, or position. However, that can have a tremendous performance and management impact on the portal. In addition, many practitioners might be assigned to a specific service line, but work on projects using a different technology or within an industry different from where they are usually assigned. The approach the portal uses to provide relevant information needs to be flexible depending upon practitioners' current needs.
The answer is to allow the end user to determine what is important by providing a palette of portlets that provide the targeted content based on service line. Users can drag and drop portlets that are important to their current needs to stay up-to-date with any particular service line (see Figure 1.12).
Figure 1.12 Using the portlet palette to manage service lines
A portlet palette is a predefined feature within WebSphere Portal that allows end users to add portlets and customized the portal pages. The use of the palette is a classic approach for meeting the needs of such a large and diverse community of users. Practitioners in the automotive sector don't always care what is going on within the energy or HR practices, for instance. Users can customize their page using the dozens of predefined portlets that target information only to those who really want it.
Another useful feature for end users looking for qualified content is the Business Research Q/A portlet. Behind this portlet is a group of research librarians who are continually mining for important data and performing analysis based on information from research sources across multiple technologies and industries. User requests to librarians are usually conducted via email; however, by integrating the Business Research portlet into the information already stored by librarians as it becomes available, it becomes much easier for end users to understand what information is readily available. Most of this information is stored in IBM Lotus Quickr and document libraries, which are fully discussed in Chapter 5, "Team Collaboration and Content Sharing." Research librarians use Quickr libraries to answer once and save data for everyone. End users can follow these links directly into the Quickr libraries to view information that has been mined by the librarians. In addition, when additional information is needed, research requests can now be initiated via the portal.
One real star of the site is the ability to run a faceted search against a number of existing repositories within the organization. This consists of a federated or simultaneous search against a number of repositories. One major problem with federated searches across multiple repositories is that each repository has its own structure and taxonomy. Therefore, search terms run against one repository might not get similar results when run against a different repository. KVLite uses a different approach that allows multiple classifications to be defined against content as it is indexed. This provides much more consistent results when running the federated search. Figure 1.13 shows the search form with multiple repositories available for the federated result.
Figure 1.13 Federated search
The search results consist of a powerful interface that integrates the results from the federated search against eight defined knowledge repositories. These results go above and beyond the provided Google-like search interface. Figure 1.14 shows some sample results that integrate sorting results by end-user ratings, tagging counts, title, author, and relevance to the search criteria.
Figure 1.14 Federated search results
In addition, the search results are integrated within a tag cloud from IBM's Dogear technology. This tag cloud allows the search results to be refined based on the tags submitted by other users of the site.
There is a lot more to the KVLite portal then we can cover here. For example, the sites tried to distinguish between searching and finding. Suppose, for example, that you are searching for some briefing documents on a particular topic. There are 21 documents available on that topic within the various search repositories. It is unlikely that you could structure a query such that you would see all 21 briefing documents in a single search result. Much work is being done to help bring related information together into drill-down types of search. Dashboards are being created that illustrate industry and technology sectors where information can be looked at rather than searched for. All this is taking place in real time while end users continue to refine the quality of the content and enhance its value to the business.
Another consideration in the Web 2.0 world is the concept of continuous beta. Because KnowledgeView Lite runs on WebSphere Portal, the team can update, add, and test new features and functionality quickly and easily. In fact, one area of the site is strictly devoted to the delivery and testing of new functionality and ideas. This idea of the constant beta is pervasive throughout IBM, as you will see in additional case studies throughout this book.