Friday, June 7, 2013

Principle #9. Complementary Information Types

Information is the blood of management systems. In regular systems there is a lot of information, but if we look at comprehensive automation, the number of information flows becomes much higher. Ideally, it would be nice to always have high quality information about everything. But is it feasible? Perhaps it could be in 100 or 1000 years, but, unfortunately, it is not feasible today. Our current technological capabilities to obtain, transmit, store, and process information are still quite limited.

Designers of management automation systems have to limit their appetites and carefully choose what information to collect, how to process it, and where to transmit and store it. It is possible, of course, to take a minimalistic approach and consider only information, which is critical for automation to function properly, and then throw away everything else. But life is not that simple. It doesn’t matter how sophisticated automation is, it can’t predict everything that could happen. At least, it can’t do it yet. At the current level of technology management systems still heavily rely on people--and the smartest man can only do so much without information. For that reason, designers have to make choices and load systems with additional information, which can be made somewhere by somebody else. To help system designers think about that problem and solve it systematically, we introduced the 9th principle: Complementary Information Types.

Now, let’s look at the information types that exist in management automation systems.

The first way to differentiate information is to look at whether the information is used by automation or not:

  • Formal - information processed by automated functions and used in management.
  • Informal - information collected and stored, but not used by automation directly. Such information can be requested by humans. And for that reason it must be supported by systems.

It is obvious that formal information is essential. Nothing can be done without it. But informal information is optional and has to be considered very carefully. If it is completely removed from a system, managers will struggle and walk in the dark. What can be worse than that? On the other hand, if it is collected without any consideration, it could bring a system’s cost up or even kill it completely. That’s why decisions have to be made consciously. You have to define what goals will be achieved by decision makers and what functions will be performed. You need to find out what could go wrong and how bad it could be. Then you have to decide what information is required for regular operations and what will be needed for emergency situations. Finally, the rest of the information that you have should be prioritized by importance, frequency of use, cost and ability to obtain, and then cut off by using the Pareto principle. From all information flows, then remove the formal part, i.e. the part required for automation itself.

The second way to differentiate information is to define it by its type or format:

  • Structured - information, presented in well-known format, which can be directly interpreted and processed by a management automation system.
  • Unstructured - information, which cannot be directly interpreted and processed. Usually it is represented as some sort free format like unstructured text, voice messages, pictures, or video.

It seems the rest is quite clear--automation needs structured information and everything else is garbage. But don’t be that quick. Structured information has a few serious drawbacks:

  • It may require extra resources to obtain.
  • While structuring, the data is simplified and abstracted. This can result in the loss of significant details.
  • Lastly, information isn’t always required to be in structured form at the beginning. Sometimes, it is sufficient for information just to be present. Its structuring can be postponed for a while.

I don’t know whether you noticed it or not, but in the previous paragraph I used the term “structuring.” This means the initial data processing, interpretation, and transformation of information into a well-known format convenient for processing. If we separate information from its structure, we can discover a few things:

  • Not all information has to be structured.
  • Not all information has to be structured at the beginning. Sometimes the structuring process can be postponed.
  • Unstructured information may contain useful details.
  • There are many methods to structure information. The most sophisticated methods use complex image-recognition algorithms. They are able to recognize information in text, speech, video, and so on. Recently, those technologies became extremely advanced and mature. But the simplest and still most widely used information structuring component is a human being. Just give someone a picture and ask him or her to describe what is seen it. You will be surprised by what he or she can find, even without seeing the image with their own eyes... Do not forget about that!

Here is the definition of Complementary Information Types:

“Management automation systems, in addition to having formal and structured information, must also support informal and unstructured information in order to respond to unpredictable real-world situations. The structuring of information in this way enables its interpretation and effectively increases the level of automation.”
Key points of Complementary Information Types:

  • Information in a system can be divided by its usage into formal and informal types, and by its ability to be interpreted into structured and unstructured types.
  • Automated management functions require formal information presented in a well-understood structured form.
  • At the same time, modern systems still rely on people and are not able to cover all possible life situations. Human beings at work need information that systems must collect, transmit, store, and present on demand.
  • Information structuring can happen at different times and in many ways. Knowing that, it is possible to choose optimal strategies by which to process information.
  • Human capabilities significantly surpass modern technologies in information structuring. Humans are able to do their work based on unstructured information. And, at the same time, they can take the role of information structuring components in automation systems.

Usage of Complementary Information Types:

  • Know about the information types in management automation systems and make conscious decisions about what will be collected, and when and how it will be processed.
  • Do not stop at the bare minimum. Consider including critical information that can be used by humans.
  • Look at the big picture beyond system automation boundaries. Define the goals of managers, their functions, and what information is necessary for them. Based on that analysis, include additional informal information.
  • Information structuring, typically, is an irreversible process. In some cases it can be delayed. Use that knowledge to improve your system.
  • There are many ways to structure information, and the technologies to do so are quickly being improved. Currently, the most advanced technologies use image recognition. Use them together with reusable components to integrate them into various systems and subsystems. Also, keep yourself informed about the latest technological advances in this area of information structuring.
  • Finally, do not forget about humans as being the most affordable, simple, and universal information structuring component available. Sometimes that route can be used to solve extremely complex problems, which still cannot be done by computers. For instance, to search for terrorists, many government agencies use specially trained people to scan innumerable phone conversations and emails.

This is the last of the 9 principles of management automation. It’s possible that this list will grow over time. But what is already there, we hope, will help you create better management automation systems today.

1 comment: