Does Technology Matter for Knowledge Management?
When knowledge management implementations fail, the blame usually falls on change management, business process or some other soft factor. After all, the proximate cause of the failure is most typically a shortage of good knowledge in the system. Why then would technology be to blame?
Knowledge management implementations face a variety of environmental, operational and cultural challenges:
Environmental—If the key challenge is capturing the knowledge of the extended enterprise, then a diversity of formats and sources is intrinsic to the problem of knowledge management. Attempts to re-engineer the way knowledge is generated in order to standardize how it is captured are generally unsuccessful.
Operational—Knowledge is constantly evolving as new solutions are identified, sometimes replacing the old solution or augmenting it. The progress of understanding is irregular and complex. What begins as a single problem can develop multiple cases—some with well-understood solutions, others with new solutions and still others that remain unsolved.
Cultural—The best knowledge is often created in response to very real and concrete challenges, for example, mapping capabilities to customer needs. Contributing knowledge to a corporate repository, while important to a knowledge manager, is simply not a high priority when faced with the next urgent task. Answers created in an artificial context just to populate a knowledge base are, by comparison, sterile and abstract.
Key Success Factors
Given the breadth of challenges facing a knowledge management implementation, teams clearly need to do some homework. Examining both successful and failed implementations yields three factors common to the successful projects:
1. Inject high-quality knowledge into the knowledgebase. Striving for quality may seem obvious but it is often hard to maintain this simple focus. IT delivers the system while users measure IT against the quality and functionality of the system. Rarely do users think to measure themselves against the quality of data they contribute. Projects are often considered nearly complete when only a handful of tests have been conducted on a small subset of the knowledge. Moreover, the format of the results is being judged, not the ability of the results to solve problems.
2. Tightly integrate with existing operational processes. Nearly every knowledge management system claims to report on knowledge effectiveness. In fact, most effectiveness data comes from user self-reporting, an unreliable method at best. Think about it: is a user likely to return to a knowledge repository and confirm that a particular document helped them complete an RFP or solve their user’s technical problem? The only way to measure knowledge effectiveness is by tightly coupling the knowledge base with core operational systems.
3. Fully leverage knowledge in its “as-created” form. Only a fraction of an enterprise’s knowledge can be formatted in any standardized way to be useful, so it is critical to leverage knowledge in whatever form it exists. The vast majority of technical know-how exists in case histories, technical documentation, white papers, discussion forums, design documents and project deliverables. If even 1% of this knowledge is harvested into a structured form for knowledge management, an organization ranks well ahead of its peers. The imperative to fully leverage the remaining 99% is clear.
Technology or Business Challenge?
To what extent does success in these three areas depend on technology? Most would assume very little, but they would be underestimating the potential for technological improvement.
Workflow solutions are the traditional approach to bringing high-quality information into the knowledge base. They can incrementally improve productivity and support a larger number of contributors, performing more tasks. But capturing even 10% of an extended enterprise’s knowledge into a structured knowledge base is still beyond reach. There is enormous potential for a technology to automate the transformation of documents into structured knowledge.
Integrating with existing operational systems, while technologically demanding, creates a secondary, important issue—how the user will interact with the system. Consider the merger of a knowledge management and a call-tracking system wherein users describe problems, are presented with solutions and request assistance. The interface is not an out-of-the-box application screen, but something heavily customized to integrate with existing call-tracking systems.Finally, the problem of leveraging knowledge in diverse formats cries out for more than a search engine. Given a large set of long, complex documents, search engines simply indicate which ones are relevant. There are two profound limitations to this approach:
- Relevance should be measured at a much lower level of granularity than the document; and
- Results should be returned at a much lower level of granularity than the document.
A system which understood the fine-grained structure of a document—its sections, chapters, paragraphs, etc.—could do dramatically better.
Conclusion
It is clear that the technology underlying a knowledge management solution is extremely important. First, the degree of automation with which disparately formatted information can be transformed into a structured knowledge base will be a critical determiner of the quality and breadth of information available. Second, the integration required with transactional systems to achieve accurate feedback on knowledge quality renders most packaged user interfaces irrelevant; however, this means the project will be largely a custom development, opening an array of technology choices. Finally, even with strong transformation capabilities, most knowledge will still reside in a variety of original document formats, in dire need of a better solution than today’s text search engines.
Max Schireson is responsible for ensuring successful customer deployment of Mark Logic products. This includes the delivery of professional services as well as the creation of value-added applications and solutions built on Mark Logic’s technology.
Mark Logic provides the first enterprise-class database for disparately structured content. Mark Logic Content Interaction Server enables enterprises to analyze, synthesize and enhance business content locked inside Microsoft Word, PDF, e-mail, Excel, PowerPoint and HTML, SGML and XML documents. Content Interaction Server provides the platform for a new class of content-centric applications. For more information, visit Mark Logic