How to... turn data into information

Knowledge may be power, but where do you go to get it? asks Annie Gurton

Most organisations have huge volumes of data in many different forms and on a variety of topics. The challenge facing IT managers is to turn that data into useful information which can be used to aid management decision making.

The arguments for converting data into knowledge are compelling. Information makes more of assets, enhances the decision-making process, saves time, reduces corporate risk, helps the business grow and improves corporate performance.

But converting raw data into information is not straightforward ? whatever the decision support software vendors tell you.

First, you have to respond promptly. If your store of data is left to grow, some of it becomes dated. Old data has a place in trend analysis, but it can corrupt the picture.

It is not enough to extract basic summaries or reports. The information has to have relevance and must be intelligent. The best applications use some kind of continually learning expert system which converts the basic data to a format which has a structure and recognises and uses processes with which managers are familiar.

The danger of basing business decisions on information extracted from raw data is that the original data may, in some way, be flawed. The user has to have confidence that the extracted knowledge is reliable and accessible, and that raw data is being interpreted in an intelligent way.

The information also has to be credible. One way to ensure credibility and to guarantee that the user has confidence in it is to explain the techniques used to glean the knowledge from data.

There needs to be some kind of audit trail of rules, so that the manager using the information can be sure that all the important elements have been taken into account when the raw data was manipulated. The rules need to be presented in a logical format which the manager easily understands and approves.

One issue, often ignored, is that the links and relationships between the sources of information can be just as important as the nuggets of information themselves. The complex flows of priorities and the history of how data is acquired can create a larger picture which can be as valuable as the information itself in the manager?s decision making.

Some experts advise the creation of a component that sits between the data and the information, to act as a buffer. This middle layer ? neither data nor information but containing both ? should be accessible to the manager seeking help with decisions. It can help to generate confidence if the information is to be used effectively.

Although the IT manager has to use technology to convert data to information, a typical user will not be concerned with technical issues. It is important to make sure that the data can be viewed from a business rather than a technical perspective, and that it is possible to interrogate the information without any understanding of the technology. This can only be achieved if the IT manager has a business bent and an understanding of business imperatives.

The software for extracting information from raw data needs to have a simple and logical interface for busy managers to use, and it must use plain English.

Delivering information derived from data requires a ?push? approach. Once the software is told the parameters of the decision, it should search the data for relevant files and then deliver information to the user.

If the manager is undertaking scenario planning, the software should be able to mine without specific commands to deliver the broadest range of ?what-if? pictures, illustrating the impact of various decisions.

Sometimes, IT managers spend many hours translating data into information, only to find that their managers ignore it. This will invariably happen if the user believes that the information has no value or cannot be trusted.

So, convincing the managers that the quality of their decisions will be enhanced by using the information is almost as important as extracting the information itself.

It is crucial that IT managers manage the construction of systems and processes to extract knowledge from data properly if the information is to be of any use.