In-Depth

The road to enlightenment

• By Charles H. Trepper
• May 31, 2001

Knowledge management seems to be slowly dying as a concept. While most organizations understand the business benefits that come from a knowledge management program, few know how to implement one. The technology has even been promoted - as recently as a year ago - as a "silver bullet" to solve all of an organization's problems. But, like CASE tools in the mid-'80s, it hasn't lived up to the hype. Despite all this, knowledge management is still an idea companies should implement.

The reality is that most organizations don't have a choice - they must create and use some kind of knowledge management function in order to survive. Companies today need to be able to continuously evolve and grow with their customer base. Personalization means more than just providing a customer with a customized Web page. It means knowing your customers well enough to send them targeted marketing information, and using data about your customers' buying patterns to detect product and pricing trends. Knowledge management also lets a company decide which vendors to use and which to drop.

Get your knowledge here!

Organizations possess enormous amounts of knowledge that is stored in their employees' brains, on paper and in computers. Yet the real issue is how a company can gather and share this intelligence in order to create a business advantage. A good knowledge management process integrates people and technology with collaborative processes to create a smarter and more competitive organization. It also lets employees use knowledge from the enterprise as a whole, not just the information found in their department.

There are two core types of knowledge: quantitative and qualitative. Quantitative knowledge includes any numerical data, such as financial figures, order quantities and customer information, such as age and income. Qualitative data includes best practices and non-numerical customer data, such as product comments by customers. A good knowledge management system lets users mine and manipulate both types of data discretely and together.

A simple, yet effective, knowledge management program can be created using a data warehouse and a collaborative/groupware tool, such as Lotus Notes or GroupWise, to share information. The key is to make the information available to decision-makers across the enterprise so that they can support their business processes.

There are three basic steps to managing a corporate knowledge base. The first step is to gather and model organizational knowledge requirements. Next, you must design and create a corporate knowledge management system using tools that store, mine and manipulate data. This system must include interfaces between the knowledge base, outside quantitative tools and transaction processing systems that supply the data to the knowledge base. The final step is ongoing support. This means meeting user needs when more data is required during the discovery process, and integrating new business units that may be created or acquired.

Step 1: What do you know?

A "knowledge analyst" first gathers and models knowledge requirements. Knowledge requirements are modeled as lists of decisions, accompanied by what users need to know to do their jobs better and faster, and what they want to know. This is different than the traditional systems analyst role. A knowledge analyst can't simply supply decision support capabilities and a bunch of data. You must also bear in mind that "needs" and "wants" are very different. A user's "needs" are simply the data that allows that person to do their job. It is a user's "wants" that provide discovery capabilities; for example, finding the correlation between buying beer and diapers.

Knowledge analysts must also supply users with data from other departments and business units across the enterprise. While this can be a major political problem, it is the heart of the knowledge management idea. Most organizations have people who are walking encyclopedias of corporate knowledge. These individuals have been with the company for a long time, and they possess valuable tips and tricks for getting things done more quickly and efficiently - knowledge everyone should possess and use. However, cross-functional knowledge bases are often difficult to create because of turf battles between department and/or business unit managers.

During the modeling process, overlaps will be found between departments and business units. These overlaps in knowledge requirements and what people already know are precisely why knowledge management is so valuable. For example, it is possible that a purchasing manager in the bakery division of a food company needs the same type of vendor information that the purchasing manager in the cereal division needs. Being able to share information helps everyone involved.

Step 2: Creating the knowledge base

Once knowledge requirements are modeled, the analyst works with designers to design and create databases that store the knowledge. Remember that there will be both quantitative and qualitative databases. Quantitative data is typically stored in data warehouses that can be accessed through online query tools, and interactive and batch reporting tools. Qualitative data can also be stored in data warehouses, but best-practice and tip information is usually best stored in groupware tools such as Lotus Notes. Most groupware tools let users quickly search text-based databases for specific terms and keywords. In turn, best-practice libraries can be categorized according to business function, customer type and so on. While this can be done in a data warehouse, it is usually a little more graceful in groupware.

The knowledge bases and transaction processing systems that supply the data must be linked. This is accomplished by using extract and load capabilities in the groupware or data warehouse tools, or via middleware. Either way, the knowledge bases must continuously pull data from transaction processing systems in order to keep the data up-to-date.

Step 3: Keep it growing

The final step is to provide ongoing support. The continuous improvement and constant re-evaluation of information needs is critical to maintaining a useful knowledge base. If data becomes stale, or tools don't grow with an organization's needs, the knowledge base quickly becomes obsolete. While users will drive continuing requirements, it is useful for the knowledge analyst to make the rounds periodically to stay current with users' needs, and to find any new requirements.

Know your future

The real benefit of a knowledge management system is the ability to share knowledge between departments and business units. Yet it is often difficult to convince management that everyone in an organization needs to be connected. In addition, cost/benefit and ROI data is difficult to calculate for knowledge management programs. You may intuitively know that more information helps you do your job better, but quantifying the cash value of that information can be next to impossible.

Cultural issues and politics kill more projects than any technical glitch ever could. The major problem with knowledge management is that organizational politicos typically bristle at the idea of sharing information because many employees still believe that knowledge is power. Wrong! Sharing knowledge is where the real power is. If you do your job better because you received information from someone in another division, you look good. If you supply knowledge to someone who does a great job, they'll remember you the next time you need help ... and you'll look good.

In order to succeed in today's global marketplace, organizations must gather quality information. Most companies, however, are overwhelmed by the quantity of data they possess because they have no easy way of organizing and mining it for day-to-day use. A good knowledge management system provides simple, elegant ways to capture, share and use information. This, in turn, helps organizations create new markets, products and services.