Media & Entertainment
Structuring Convergence
Paid Content

Information
Knowledge Management

Strategy
Time Based Competition
The "Vision" Thing

Internet
The Net Era
The Internet - Some Perspectives
Chat Society
Net Governance

Branding
Brand Positioning
Brand Personality
Brand Extension

Written 1999

Chase Manhattan (Relationship Management System), KPMG (Kworld - consulting practices), Eli Lily (Pharmaceutical research and marketing) are three companies with a worldwide presence who have reaped handsome rewards from investing in knowledge. Knowledge Management has become a corporate catchphrase today. However few understand it and fewer still agree on what it involves. In this paper we will try and create a shared understanding of this concept of Knowledge and then move on to exploring how it can be managed.

What exactly is "knowledge"?
And how does it differ from information?

Context and relevance
Lets first take a step back and distinguish between data and information. The difference between data and information is context. "The score is five nil" is data as long as I have no context to the statement. However, if I know that it is a hockey match, played between India and Pakistan, there's suddenly a context to it. (Of course I would like to know who's five and who's zero, but not knowing that doesn't prevent it from being information.) The context may be very minimal. Just knowing that it's a hockey match attaches some context to it. The context is therefore dependent on who's using the data or information. For a logistics manager - "impending truck strike in Madhya Pradesh" may be contextual. For a sales manager in Tamil Nadu, the same sentence may have little context. There is a whiff of relevance, therefore in this definition of context.

Patterns
Knowledge, in this continuum, is information with patterns - patterns that are meaningful and can be the basis for actions, forecasts and predictive decisions. While 30 cm of rain in Jan may be information, knowing that Jan usually has 5% of the years rains, can be used to create assumptions about how much it will rain in the year. Analysts in the stock markets tend to use a lot of this kind of predictive knowledge.

Experience - a useful and easy definition.
Another way of defining knowledge is that it is the accumulated learning gleaned from experience. For the purpose of the rest of this note, lets stick to this second definition, simply because its easier to grapple with. A simple example will demonstrate the congruence of the two definitions. It will also highlight the difference between information and knowledge.

Kasparov vs Deep Blue
Consider the famous instance of Kasparov playing chess against Deep Blue - the IBM computer. What is the difference between the two? Without getting into the history of how Deep Blue was programmed, its clear that it can and does evaluate a few thousand times more permutations of moves, per second, than Kasparov. The computer tends to process information. It analyses all possible moves and their ramifications up to a finite number of moves. It does this each time it makes a move. Does Kasparov do this? Simple, he can't. What he does, is look for patterns. He eliminates a large number of moves and possibilities by basic pattern recognition. He evaluates only a select few options, therefore and is able to match the computer by intelligently reducing the scope of the sheer volume of processing.

Notice that this pattern recognition is borne of the years and years of experience and the thousands of games that Gary Kasparov has played since he was able to distinguish black from white. It therefore is another way of saying that he now has enough accumulated experience behind him to have learnt a whole lot that he may or may not have actually explicitly "categorised" in his mind. Knowledge often therefore takes the form of "gut feel".

Why is this pattern recognition important?
In business, most often we are forced into working with incomplete information. Or information that is constantly changing. Sometimes the time it takes for us to collect, validate and process information is more than the time it takes for the realities of the business, the technologies and the markets to change. This is especially true in the world of the Internet. In such environments, it is perhaps folly to depend on information processing as the primary decision support mechanism. It leads to incorrect decisions, made on the basis of erroneous conclusions drawn from incomplete or inaccurate information. This is why need to work with a model better than just information processing. (Of course, information engines like ERP packages from SAP tend to be able to provide information that is both accurate and timely, but even SAP's current thinking is leading to the exploration of the Knowledge Domain, rather than focusing of faster and more information.) Remember, management is not an exact science (if it is a science at all!) A good marketing plan does not guarantee that the customers will come running; there are too many factors involved. However, it would surely improve the chances of closing more orders. In the same way, knowledge management increases the chances of knowledge being captured and shared and creating the learning organisation.

What constitutes knowledge: An Example at Planetasia.com
For Planetasia.com, which is an Internet Business Solutions Company, the broad knowledge areas are, of course, the areas of Business Processes, Technologies, Design and Project management. Each of these has many sub-categories, leading to a rather complicated knowledge map, but for the sake of simplicity, lets just look at one of the areas. Within design, lets look at information design. This is a little known area, with only a handful of people across the world speaking about it. Yet, we find that this is a crucial discipline for better solutions, whether they be intranet, e-commerce, or website based solutions. This means that we need to create, study and recognise patterns in information structuring, drawing from all the work we do, and often work not done by us, or even from sources not related to the internet. For example, a lot of learning about information design can come from the field of retailing - specifically, from visual merchandising. Library sciences also have a lot to contribute.

At this point, it would be both impolitic (for me) and even boring (for you) to list all the areas which we are trying to build knowledge in. Instead, lets move to the creation of a system that can do the job of managing knowledge.

How do we manage this knowledge?

The need for Organisation Learning
First, lets state the obvious, organisations need organisational learning. It does no good for my company if I build up a lot of knowledge, pattern recognition ability and then, someday, leave, taking my knowledge with me. Worse, that others in the same organisation may be making the same mistakes, going through the same, slow, experiential process of learning.

The need for organisational leverage
Second, organisations need to leverage knowledge. This implies that in a logistics solutions companies, if out of experience, there is some learning about how to put a larger number of crates into a truck or to reduce the breakage in transport, the organisation needs to leverage this learning into more business, more value for its customers and more profits. The company always runs the danger of such learning getting localised (the Mumbai branch knows, but the Chennai guys don't).

Both these points illustrate the need for a way to share this knowledge. This necessarily implies a need for explicitising what is gut felt, implicit and intuitive. Therefore, a Knowledge Management System needs to provide a way of converting implicit knowledge into explicit knowledge and then a means for sharing this.

Not the same as Individual Learning
Next, lets understand, that organisational learning and memory is different from individual learning, and depends on systems thinking. (Peter Senge's seminal work on learning organisations is perhaps the best reference on this.) The time-span over which knowledge is built and used in organisations is often beyond typical human planning horizons. Over such long periods, there is a high risk of obsolescence of the Knowledge. Which means that it needs periodic validation and updation. So the Knowledge Management System must also enable storage of Knowledge and the Means of Updating, adding, modifying and revalidating knowledge. Lets also remember that a lot of times, gut-feel may need to be validated, and may prove to be wrong.

So how does one build a Knowledge Management System?

There is always a danger of reading "system" as "computer system". First, you need to have a logical system that can support all your knowledge management requirements. Lets look at the key aspects of this logical system:

Every Knowledge area needs an owner
The notion of democratic knowledge development is nice.
In the absence of an owner, the weeds can grow quickly all over the garden, and soon become a nightmare for anyone who wanders into it. This implies that for the first few months, until some maturity is achieved across the organisation with respect to the knowledge building process, one person is assigned as the owner of a particular area of knowledge. For example, one person would own, at Planetasia.com, the area of information design, as far as the Knowledge Management System is concerned.

Knowledge means people access
Knowledge needs to be associated with people - else it can become information. And you need to be able to access these people.
On the World Wide Web, its possible to find tons of documents about almost every area of knowledge. So what's the big advantage of creating your own little model of the WWW? It becomes a whole lot more powerful when you can attach a person to every piece of learning and make this person accessible. If a Planetasia.com consultant has done an information design exercise for a bookshop in Mumbai, then another person wishing to do a similar exercise for a retail outlet, say in Chennai, should be able to access the learnings of his/ her Mumbai colleague as well as ask the Mumbai consultant some additional questions. The power of this grows exponentially with the size of the organisation, the geographical reach and the role of intellectual property in the value proposition.

Extraction is the key to explicit knowledge
You need to have a method for extracting from the learning to create explicit knowledge.
So I've done a whole project. I've learnt a lot. How do you want me to put it down? Where do I begin? With the scheduling of activities?? Or how I learnt to create project plans for multiple projects? How I learnt to handle impatient clients? Or about how I managed an outsourced component? See the problem? Someone has to define a "template" of learning documentation. This may Sometimes mean making a choice between areas, but that's a necessary cost for better Knowledge Management. At the end of the day, you can leave it unstructured, but retrieval and therefore use is tied in to the level of structuring. The problem is much larger for bigger stores of such learnings.

Knowledge needs to be organic
To be really powerful, knowledge must be able to grow, morph and rejuvenate, guided by the invisible hand of knowledge economics.
Creating libraries of static documents is not the objective of a Knowledge Management System. So you must also create a system that allows people to add to existing knowledge, challenge earlier hypotheses, contradict earlier learning or to add to it.

What are the tools of knowledge management?

Access
Now we get into the nuts and bolts. Clearly you need all people to have access to the Knowledge Management System - so you either need a WAN, Public Network (Internet), or VPN access to your repository. This repository will typically reside in a server or a set of servers, which will be co-located or distributed.

Environment
Next you need a collaborative publishing environment. The Web makes a lot of sense, here. All the needs of this system fit rather well onto the capabilities of Net Technologies. What we are talking about is, in current business parlance, an intranet. A private, shared, collaborative, secure, publishing environment that allows you to do all the things we've spoken about till now. Within the canvas of Net technologies, you may still need to choose between environments such as Microsoft, Netsape/Unix, Oracle, Lotus Notes and IBM. Each comes with specific advantages and disadvantages, and flavours that may suit your specific needs.

Applications
Finally you need a set of applications that help you manage the knowledge. Applications for entering new content, applications for modifying, editing, deleting, integrating with mailing systems (which can enable you to directly contact the author of a document) and directory services based on LDAP standards (which will allow you to map all organisational resources). Some of the other technologies that are closely linked with Knowledge Management Systems are databases and data warehouses, advanced search tools, push technology, agent technology, brokers, collaboration, document management, help desk technologies and brainstorming applications.

Some advanced search methods
· Natural Language Searching
· Boolean Searching
· Automatic root expansion
· Proximity Searching
· Numeric Searching
· Term Weighted searching
· Thesaurus Integration
· Search by Object type
· Search by Metadata fields
· Concept searching
· Image searching

Approaches to a Knowledge Management System

There are broadly 2 schools of thought. The first approach, which can be described as "lets-define-knowledge-as-the-stuff-inside-people's-heads" approach, involves distinguishing knowledge from information or data. This usually gets down to defining knowledge as your organisation's intellectual property and does not, for example, see an industry analysis done by someone else as knowledge. The second approach, best described as an "I-don't-care-how-you-define-it; if-its-useful-and-relevant-its-knowledge" approach, considers only the yardstick of relevance and usefulness. In this case an industry report created by some other organisation would be considered as knowledge. The first approach is better suited to those organisations that are already effectively managing their data and information. This includes organisations with ERP solutions, intranets or other enterprisewide solutions. The second approach works better for companies that have no information or data management system to begin with.

Clearly, your knowledge management initiative needs to be both top down (for resources and budgets) and bottom up (for enthusiasm and usage).

Don't forget the culture issues:

Thomas Davenport, an acknowledged authority on Knowledge Management Systems speaks eloquently about the politics of knowledge. Often it is this knowledge that gives an employee his/ her standing in the organisation. Giving this away could represent a submission of power. This can prove to be a barrier to the best designed Knowledge Management Systems.

Individualism & Entrepreneurial competition are ingrained as traits within most of us. The solution is to evolve mature incentives to draw out that knowledge without being threatening to the owners of the knowledge.

The Lotus notes example: a CIO Magazine Article recounts that even within Lotus Corporation, arguably the leaders in the knowledge tools market, a tendency to ignore the culture issues led to the knowledge management process falling short of expectations.

Armed with all of the above, you are now ready to sally forth onto the ocean of knowledge. Just a few words of warning. There are no absolute truths in this emerging area. Be prepared to relearn everything you know (or have read in this note).