Efficiency of information system


Dzuba S..,

Doctor of Science,

Irkutsk State Technical University,


Usually when we talk about evolution of a company we link progressive nature of this evolution to growth of the company. In this case a company can effectively use the economies of scale which decrease the marginal costs as unification and standardization can make operations automatic. This relates not only to the production functions but also the routine procedures of calculation and control. No wonder that information system (IS), as a tool for their automatization, is expected to bring the same advantages as those that can be provided, for example, by a textile machine or a steam-engine. But it is long time since it was found that economy of management labour as expected direct effect often manifests itself rather weakly in contrast to implicit effect on general improvement of management. The latter can not be assessed consistently. Nevertheless according to all methodic guidelines for evaluation of efficiency of IS it is recommended to calculate economical benefits and compare them with costs of implementation and usage of this IS. Absence of any other methodic alternatives makes us to turn to this method over and over again, though it is based on input data which can not be reliable from the very beginning of calculation.

Lets have another look on this problem. Will any company analyze necessity of providing office with telephonic communication or evaluate economical efficiency of such decision? If it is not bureaucratic to very high extent it will not. Need for telephones is obvious without any analysis and the benefits obtained from telephonic system exceed the costs of its installation and maintenance endlessly providing that personnels work is connected with constant message exchange between each other and third parties. Can we say that any way of organization of telephonic communications will be equally effective? No, we can not. Let us consider three typical schemes:

  • Every employee has its own telephone number. This variant is the most expensive because the costs directly relate to number capacity (the quantity of numbers).
  • Every work group has its own phone number.
  • Phone communication inside the office is provided by the office phone station which has one multi-channel city phone number. If extension numbers (for communication inside the office) are much cheaper than city numbers this variant is the most cost-effective.

The first variant looks like the most luxurious. But the employees will have to memorize very many numbers and constantly use indoor telephone directory. If the necessary employee is not at his place a caller which needs to transmit the message will have to wait until he comes back or to make another call to his neighbor (a caller needs to know this number or make enquiry about it). The employees neighbor may not be at his place as well. And so on. In the third variant it will take a lot of time and efforts to make a call to the necessary employee because a caller will have to listen to the music and every time explain to a receptionist who he is and what he needs.
It does not make sense to clarify which variant is the best. What is more important for us is that we can correctly compare them without taking the value of economic benefit into consideration. The same principle can be applied to IS in general. But to make this hypothesis more convincing we shall have to take a very long way.

The existence of costs of market regulation which classic economic theory does not take into account was mentioned by Ronald Coase in the work The Nature of the Firm. He started to take into account transaction costs which can be seen as economical friction force originated in the process of coordination of interaction circumstances between economic agents. Companies exist because the volume of such coordination inside them is lower than in external market environment. But this can not last all the time: in this case the companies as a way to allocate economical resources inevitably would have taken the place of the market but it does not happen in reality. The reason for that, as R. Coase mentioned, prohibition-like forward-looking increase in transaction costs inside the company which follow the growth of the company. Further the prohibition of Coase takes place which means stopping of company growth as a result. Coase did not explain the principle of its action but he pointed out that it can be smoothened by improvement of management of the company.
If a company can be seen as cybernetic system we can observe that a well-managed company can reduce the variety of its states which means that it can put them in order. The ability to do it is described by the Ashbys law of requisite variety. It says that variety of impacts on the system can be reduced only by the value of variety of management, not more.

It is rather strange that in reference internet resources this law was transformed into very alike and similar in sounding but absolutely false in its contents statement which is sometimes cited as Ashbys words: To make the management of the system possible the variety of managing actions must not be less than variety of agitation at the entrance into the system. Of course Ashby could not say such nonsense and actually did not because in such a form the law let the possibility of achieving the negative variety by the system and this breaks the second law of thermodynamics (principle of entropy increase).
However even in the correct form the law looks like some banality because it only says that complex systems must have complex management. If it is not possible to achieve necessary complexity of management the system should be broken down into smaller blocks organized in a hierarchical way. This was the understanding of the statement accepted by the scientists thanks largely to the author himself as Ashby believed that managing sub-system (regulator) was fully trained. His favorite illustration was a fencer who in order to avoid destabilizing his inner sub-system as a result of, for example, his rivals successive attack must defend in maximal diversified way. But this indirectly suggests that the regulator can find specially developed action for every agitation of the system which transforms the system into necessary state. Otherwise the fencer can act in a maximal diversified way but if its responses are not strictly determined by the rivals actions he will have little chances to fulfill his task.
It is understood that the companys leader (or the company management) is obviously a trained regulator. In reality it is not quite so or not so at all. Even the most skilful leader who comes to manage a new company needs some time to be spent before he clarifies the situation in the company. But this is not the only problem. Processes of training and de-training of the pair company-management take place all the time. If the first process is quite clear for the reader, the reason for de-training is transformation of system being managed or changing of set of its states. The simplest feature of this set will be the number of its elements, in particular it is general quantity of its states. It is clear that this quantity grows as the company gets bigger. This results in new states being added to the set of states. This leads to the de-training of the regulator which grows exponentially, i.e. is of prohibition nature (prohibition of Coase). Management system of a company and IS as one of its tools can decrease the value of this exponential curve making it more flat but they can not eliminate the tendency of loss of manageability completely. As the main result of functioning of IS is smoothing of the Coase prohibition we can say if it is effective or not if we compare its efficiency and costs.

These statements as well as the methods for evaluation of efficiency of IS and their theoretical base are represented more rigorously in the articles Efficiency of management system: information approach and Efficiency and costs of information system: theory and practice published in the magazine Management in Russia and Abroad.

«» - «».