Most knowledge management systems go on
life-support during the first two years after implementation.  Management wonders if they will ever really see measurable, long-term business value.  While most users resort to other means to get to information, a few zealots continue the quest to really make it work.  Then after three years, the business decides to pull the plug and start over—either by buying replacement technology or re-launching the existing technology.  What your knowledge does not know may ultimately destroy it.  For example:

• When content is created too far from or close to the point of demand, it loses perspective resulting in too general or too specific level content.  Because the content is not quite right, it causes a need for more content resulting in partial and conflicting content.  Too much content is “fat” which will clog the system. The knowledge system must know DEMAND context.



• If the knowledge system is meant to address logic (i.e. troubleshooting) as well as retrieve information, the structure of the situation must guide the user through resolution.  If the logic is built into the content itself, too much content is needed to fit too many variations of the situation.  Ultimately, some content will work for experts and other content will work for novice users causing the overall set of content to be confusing and misleading.  The knowledge system must know RESOLUTION context.



• If the content relies primarily on search terms and vocabulary for establishing relevance, eventually the same terminologies will be found in too many places.  The users’ perspectives will continue to vary causing the need for broader vocabularies.  Eventually, the vocabulary becomes too homogeneous and the search engine is unable to differentiate relevance.  The knowledge system must know ACCESS context.

Without this level of knowing, the knowledge system becomes dysfunctional and although the technologies still works the content becomes stale and a burden to the organization.

To address some of these issues, the Knowledge NormalizationSM method was created.  Normalization is a method which structures group knowledge into findable, context-driven solutions.  The methodology enables engineers or technicians to solve more problems without escalating.  It also enables customers to address their own issues before initiating support requests.  Below are the three aspects of Knowledge Normalization we focus on to keep knowledge Fat-free, Fit, and Flexible.

• Demand Analysis — supports fat-free content



• Resolution Paths — content fits complex, varied situations



• Access Topology — gives flexibility to get to relevant content results without relying on consistent searches

Fat-free:  Which content is muscle (i.e. an asset) and which is fat (i.e. a liability)?

Some content gets in the way and some gets at the user’s need directly.  It becomes difficult to isolate which content is which.  Both types of content use the same terminology.  And both types of content follow the quality guidelines outlined in the content standard.  Duplicate content is not only two documents which have the exact same content, it can also be two different answers to the same question. You cannot tell good content from bad by simply looking at the content itself.  You can only tell good content from bad by the way it addresses specific user demand.

In this instance, demand reflects any type of support interaction (i.e. a logged case, a web interaction, a chat session or an email).  Content is only “good” if it successfully addresses demand.  The hit rate on content does not provide an accurate indication of demand.  By conducting an analysis quantifying key attributes of support requests, you can organize support requests demand into prioritized buckets.  The analysis quantifies the value of support requests (i.e. according to customer, product and severity of issue values) and contextualizes the demand against the user experience (i.e. what is the user intent). The buckets are allocated to subject matter experts who determine which content works and which content does not work for a specific user intent. The demand analysis sets business value for the content and relates it to customer needs.  With this method, all content stays in shape and content “fat” does not get created in the first place, resulting in a lot less work. 

Fit: Is the content fit for complex situations?

It is easy for quick answers or comprehensive documents to be retrieved effectively in a search. It is not easy when mitigating conditions make many possible courses of actions the right answers.  Such is the case in complex situations.  If the alternate course of answers have serious consequences, the ability to quickly and accurately move toward the right answer becomes critical.  The right answer depends on how it relates to the situation in many ways.  Decision trees are sometimes used to support complex situations.  But they are often tedious to manage and too rigid to use.  The user experience suffers when the big picture of how one step relates to another is not exposed.  The logic to be used cannot be “hard coded” because “it depends.” 

The logic of the situation must be implicit in the relationship between content objects — with each object acting as a step in path.  That relationship must be managed by experts to guide less expert users toward the right answer.  Resolution paths are one method to support this requirement.  They are created to organize content objects into a path of logic — often using the same object in multiple points across different paths.  When experts have a landscape view of the logic applied to situations and the visibility to see which objects fit on which paths, they enable the business to know which problems drive demand.  The business will know how long each type of problem should take to solve.  By understanding the path a specific type of problem takes, experts improve the efficiency of resolution overtly.  Then creating content becomes intriguing, not a chore or a hygiene item. Improved efficiency in solving complex issues becomes measurable quickly. 

Flexible:  Ability to stretch across search and navigation no matter how weak the search string

Often the user does not know what to look for. Even when they do, they express the search in many varied ways. The demand structure, not just the search terms, strengthens the relevance of the content for the user.  The user has a finite number of things they want to do with their technical environment.  However, the numbers of issues they can experience with the technology are infinite. If the user has a consistent and simple navigation architecture - based on their intent, rather than their problem - they will be able to find what they need without having to explicitly express it. The Knowledge Normalization process produces an organizing structure called an Access Topology to make content easier to navigate than the categories of possible problems. The Access Topology reflects a customer’s perspective of a product. Using a 24-node taxonomy model, any area of support can organize its content so the user, at any level of expertise, can find it in three clicks.

The Knowledge Normalization method has been used by many companies (i.e. EMC, Juniper Networks, Legato — A Division of EMC, Sun Microsystems, and VeriSign).  They see results quickly (up to 53% reduction in time to resolve within 3 months of publishing normalized content) because the content is aligned to demand (fat-free), is highly relevant (fit) and easy to find (flexible).  The ability to maintain the value of the content long-term is sustained through the healthier organizing structure.

About Livia Wilson…………………………………………………………….

Livia Wilson is the president and co-founder of Outsights, Inc. Livia has more than 20 years professional experience in organizational leadership and development, knowledge and program management, software and systems design, and strategic support of service delivery practices.  Outsights supports implementation models for knowledge-centered strategies.

Livia can be reached at lwilson@outsights.com or 772-539-9378. Visit Outsight’s website at www.outsights.com.