Summary

• Knowledge is re-usable.
  
  
• Knowledge is different from information. Both are needed for effective decision making.
  
  
• Knowledge of a domain helps identify which information is relevant.
  
  
• Strategic Knowledge Management addresses the four major weaknesses in the strategic decision making process.
  
  
• Strategic Knowledge Tools require explanatory power. Fuzzy logic provides this.
  
  
• A worked example is provided on a pricing decision. It uses the Assistum knowledge tool.

1. Knowledge -  What is it?

A successful business makes good decisions, implements them well - and then learns from the experience in order to do better next time.

To make a good decision, one needs not only information about the specific instance, but also an understanding of the domain. In other words, one needs a set of principles, models, templates or other abstractions.

These abstractions are then re-usable for making new decisions with different information. Knowledge is re-usable - unlike information which relates to specific instances.

Knowledge is a set of re-usable abstractions that assist understanding and provide meaning to decision-making.
Information is about specific instances and is the raw material of particular decisions.

For instance: Consider a civil engineering business that builds bridges.

Its knowledge is in its understanding of how to build bridges, how to manage projects, how to handle finance etc. This knowledge is re-usable.

Its information is about specific bridges, budgets, suppliers etc which are essentially transitory.

If it knows how to build bridges but has no information about particular bridges that need building, it is not going to be in business for long!
Conversely if it has a lot of information about particular bridges but no understanding of how to build a bridge, it will soon be out of business.
It needs the combination of knowledge about bridge-building AND information on bridges.

Knowledge

Re-usable abstractions

equations

principles

rules

insights

recipes

models

templates

diagrams

etc.

The basis of human understanding

2. Knowledge and Information – They are different!

How often have you heard Boards of Directors defer a difficult decision and ask for more information? The reason, in many cases, is that they inadequately understand the key decision drivers and how they impact the decision. They seek comfort in asking for more information.
Managers often mistakenly seek more and more information instead of better knowledge or understanding.

Better understanding enables the identification of what information is relevant. Consequently, less information is required because the irrelevant can be ignored.

Better understanding, at the conventional "best practice" level, can be the springboard to innovation. Could Einstein have developed the theory of relativity without an understanding of standard best practice, i.e. Newton’s Laws of Motion? It would have been much more difficult.

The challenge is to encourage the capturing and understanding of business best practice at the strategic level to improve performance and provide the platform for dramatic innovation.

3. Strategic Knowledge Management – Making the big decisions

Business decisions come in all shapes and sizes - small and large. The small I will call tactical or operational; the large I will call strategic.

Operational decisions are typically many and small with sufficient prior data on which to build a knowledge base. This is the field of data-mining and neural networks. Knowledge based systems for tactical decision making can be automated taking human understanding out of the loop except at the boundaries e.g. credit scoring.

Strategic decisions are typically infrequent, each of large value with little formal prior data of similar decisions to build understanding on, and requiring business judgement, and the balancing of qualitative trade-offs. Making a business acquisition is an example.

Knowledge-based systems that help in strategic decision-making assist the human decision-maker by enhancing her understanding. This is the subject of this paper.

4. Weaknesses in the decision-making process

There are four potential weaknesses in the strategic decision making process.

1. The process of obtaining relevant information.
   
   
2. The process of re-using existing knowledge and avoiding the NIH (Not Invented Here) problem.
   
   
3. The process of making good decisions quickly – i.e. not making bad decisions quickly nor avoiding making any decision at all.
   
   
4. Having made a decision, the process of turning the decision into action.

This paper addresses all these weaknesses.

5. Fuzzy logic and explanatory systems

The knowledge used in making strategic decisions are often expressed in words as "business principles" or "rules of thumb".

If you listen to business people discussing a strategic decision you will often hear these "rules of thumb".

"If a customer is wealthy, they are likely to be less price sensitive"

The source is either from their own business experience, or from business consultants or business schools, or the latest management books or gurus.

You will recognise these "business rules" as fuzzy rules with fuzzy states such as "poor", and fuzzy impacts such as "likely to be".

These fuzzy rules can be combined using fuzzy logic.

The particular application I am about to describe uses the ASSISTUM knowledge tool.

6. System requirements

The key for strategic decision making is that the knowledge systems should be:

• Capable of explaining its reasoning to business people
  (rather than incomprehensible lines of jargon or code)

• Explicitly under the control of human users
  (to enhance their understanding, not replace it)

• Fun and stimulating to use
  (or it will not be used)

• Easily designed and modified
  (to incorporate new learnings and insights based on its use in practice

7. Applications in strategic decision making – an example.

Knowledge based systems of this type have been used in many strategic areas including pricing, outsourcing and portfolio management. The example I shall use is a simplified version of a pricing knowledge base.

It is worth noting that because the human user is in charge and makes all the decisions, the system does not have to be mathematically provable or verifiable, - just sufficiently credible and stimulating that it encourages human users to explore and share their own understanding of a domain.

The system performs the same function as a management book or a business consultant, neither of which are required to be mathematically provable. However it is better than a book, and much cheaper than a consultant!

The knowledge base displays the issues or concepts that need to be considered when making the final decision - in this case "what is the potential for increasing the price of this product in this market?".

The knowledge base has business rules (in this case on pricing issues) but it currently has no information about the specific scenario you are about to investigate (in this example the case for increasing the price of tomato soup).

It is important that the investigation is done by people with some experience of the specific area under investigation. Although the knowledge base contains general business knowledge (in this case about pricing) it contains no information about the specific scenario.

I recommend the use of an experienced team with different perspectives of the area (marketing, financial, manufacturing etc.). ASSISTUM will then help the team pool its experience to make a better decision. It does not replace them or their responsibility for the final decision.

We will investigate customer wealth by double-clicking on that line.

The value of ASSISTUM lies in the quality of the conversations it provokes and new insights it stimulates. In this case, the conversation will be around the target customers or "end-users" and how wealthy they are. After a heated debate, we agree that though some of our customers for tomato soup are very wealthy, the majority are of below average wealth and we chose "fairly poor" .

We answer "very old" to age of product, and "very frequent" to frequency of purchase. This leaves us with the screen above.

After a short discussion on what this means for customer familiarity with the price of tomato soup, we press Proceed.

Before ASSISTUM will divulge its conclusions, it requires the team to first come up with its view. This is to stimulate discussion of the issues and get some ownership of the conclusions.

In this example, the team decides that customers are "fairly familiar" with the price of tomato soup.

The knowledge base disagrees with the team and gives its reasoning.

The team must now decide whether to accept the ASSISTUM conclusion, reject some of Assistum’s line of reasoning or continue with their own because they know of some other factors.

The human team must finally decide. This is key to the value and integrity of the ASSISTUM method.

In this case, the team accepts the ASSISTUM conclusion.

We then continue in this fashion investigating all the issues.

Finally a conclusion is reached.

We agree that a price increase is probably indicated after all, in spite of the customers’ price sensitivity.

It then displays all the conclusions and the way they are related. Green links support a price increase. Red links deny it. The thickness of the lines show the importance of the impact.

It is also possible to display the conclusions in report form for those who prefer to read words rather than look at pictures.

I hope this has demonstrated that it is possible to meet the system requirements I earlier.

8. Knowledge and Innovation

There is a cycle to the creation and use of knowledge. We have already mentioned how increased understanding of current best practice in strategic thinking can lead to innovative breakthroughs.

Current knowledge is explicitly represented. It is processed by human brains (together with information about the specific issue at hand) leading to understanding, decisions and action.

This is sometimes accompanied by new ideas or hypotheses. If supported, the new ideas lead to innovative actions, and also to new knowledge that is added to the explicit representation of current knowledge.


Ian Lang
Higher Level Systems Ltd               Assistum Home page                Details of products
June 1999

For any further information please e-mail us at Assistum