Technique 6.4: Closed Loop Feedback for the Whole System

Root Cause Analysis and Evolve Loop Content

Just as the healthcare industry has moved from reactive-only services to more proactive, preventive care over the last decade, many organizations have started to supplement reactive support with preventative actions—eliminating the source of issues in the first place. This has become possible as self-service has off-loaded responders' time, making them available to spend more time identifying issues for elimination. In order to find and diagnose problems, Knowledge Domain Experts perform root cause analysis. The patterns and trends of the articles in the knowledge base are the source of information for the analysis.

Knowledge articles are very transactional in nature. They represent what we have learned from an interaction. Looking at a collection of articles in a domain allows us to identify patterns and trends. We can assess the closeness or distance between articles. Articles that cluster around common themes or have similar causes represent opportunities to improve products or services. Removing the source of a frequent request is the ultimate level of success for an organization as it improves the customer productivity.

80% of the incident volume relates to 20% of the content in the knowledge base.

The Knowledge Domain Expert, product engineering, and product management must be part of the workflow and become engaged as patterns and trends in the Solve Loop content start to emerge. Through understanding the patterns and trends, we can pursue another form of intervention. Perhaps we can improve the documentation or create Evolve Loop articles - ones that merge the experience represented in many related KCS articles into a single KCS article. This idea is discussed further under the role of the Knowledge Domain Expert and in the Creating Evolve Loop Articles technique. (See KCS Roles and the Licensing Model for a complete description of Knowledge Domain Expert responsibilities.)

Continuous Improvement of the Workflow and Content Standard

The closed loop nature of the workflow makes it easier to monitor and maintain its effectiveness. This is done by continually sharing best practices. Gathering the best practice feedback must be part of the continuous improvement process. Because KCS articles are effective for procedural information, many organizations use their knowledge base to collect information about the KCS workflow or the content standard. Questions and issues about the workflow, the tools, or the content standard are captured as articles in the knowledge base (often marked as internal use only).

For example, if we are seeing many incomplete or duplicate articles, then we most likely have a workflow problem. The most common cause of duplicate KCS articles is a breakdown in the problem solving process; people are not searching before they create (search early, search often). Here again it is the role of the Coaches and Knowledge Domain Experts to investigate these kinds of issues and work to improve the level of understanding of the KCS practices across the organization. We should also mention, relative to this example, that management can be the source of unwanted behaviors by putting the wrong metrics and goals in place (more on this in the Performance Assessment section), or by not revoking the KCS license when an individual's performance slips below acceptable levels. A system of continuous improvement is critical to sustain the practices and must involve the Coaches, Knowledge Domain Experts, and managers.

Early in the adoption phase of KCS, the Coaches play a critical role in refining the workflow. As the adoption matures, the organization needs to establish a way to manage a continuous improvement process. This is often done through a KCS Council (see Performance Assessment for descriptions of these roles and responsibilities).

This continuous improvement can affect both the process and the content quality itself. For example, a software company may ask a Knowledge Domain Expert to focus on using knowledge management tools to monitor search strings, KCS articles found and considered, and the incident record to understand the KCS article process. With this insight, the Knowledge Domain Expert will be able to create high value Evolve Loop content.

Variations on a Theme

Process is important, but be smart about it. Success with KCS requires an understanding of the concepts such that the practices can be tuned to meet the needs of a specific environment. For example, what if we have an environment where we experience an extremely high frequency of a few problems or questions? Articles for these problems exist in the knowledge base. Normally we would expect those articles would make their way quickly to the web due to the high reuse.

But what if the environment is such that web delivery is not an option, or just not ready yet? In this case, does the full process of search, capture, and structure make sense for these few frequently raised issues? Since the responders answer these issues many times a day, they know the answer, and they know it is in the knowledge base. It does not make sense to force the standard KCS workflow. In this situation, the standard workflow offers no value.

For these frequently raised issues, we should enable an easy way for the responders to record that they answered this question again. This shortcut is often a "quick click," or a favorites list that allows them to quickly record the reuse of the KCS article. For high volume issues the important thing is to capture how often the KCS article is being used, so the "quick click" button needs to drive the reuse counter for that KCS article.

This list of frequently provided answers may be unique to any group of responders. There should not be more than ten items in this list. Also, our processes should include ways to update the responders if the preferred answer changes over time. For example, if a better answer evolves, when the user hits the "quick click" button for that KCS article, the new information displays to update the responder.