Economics of information systems and urbanization

(en français)

(Conference " Urbanization of information systems ", Paris Sorbonne, June 24, 2002)

To evoke in a few words the economics of information systems we will define the service which it renders to the firm, then speak about its cost and profitability. We will conclude by some simple rules of the thumb for the managers. 

Information System:  a Language

What is the service which the information system provides to the firm?  here the most relevant definition:  the information system is the language of the firm (not a programming language, but the language which is spoken in the firm). The referential defines the concepts used by the firm for describing its customers, products, organization and procedures.  This language is moreover equipped by an ubiquitous programmable automat  which assists the mental work of the user by providing means of classification, sorting and processing data.

Data processing was in the Seventies aimed at structured data only;  it concerns today the whole process of design and production of the firm. It computes and classifies texts in natural language, mail, documentation, in short everything that is written in the firm. Only the erasable texts, oral conversations and handwritten notes escape to him.  The management of confidentiality becomes essential:  the information system equips it too. 

If the information system is well conceived, the firm enjoys a good language to clarify its positioning, to make its process work and to evaluate its action.  It controls the multi-media communication between its agents and with its customers, as well as interoperability with the information systems of its partners. Then the information system elucidates  the firm:  it enlightens it so that it radiates the information necessary to the actors.  One hears in the firm sentences like:  " one knows what one has to do ", " everything is well organized ", " the firm goes well ", " the management is efficient ". 

The information system was often defined not by its aims, but by the means  it uses : computers, programs and networks.  It was believed that information system equals data processing.  But the design of the language of the firm concerns the responsibility of the business units.

Quality of language raises delicate philosophical questions; more, structuring and equipping a language results in defining an organization. Hence some blockages which are simultaneously intellectual and sociological. The commonplace " business is business " does not allow to see there clearly.  Our firms advance slowly and awkwardly, moving back as a person who would be pushed by a hand posed on her chest and which would stumble on obstacles that she cannot see. 

An Asset whose Cost is Badly Controlled

The information system is for the firm an asset,  a productive equipment, even if it is of a very peculiar type. The applications form a portfolio ; it is necessary to manage him as well as the other components of the balance.

However accounting and tax conventions mask this reality.  The expenditure of development being regarded as operating costs, the information system does not appear in the balance account.  The AFAI1   builds a nomenclature which will make possible to evaluate the cost function of the information system, but up to now it did not exist

The legal constraints of the tax systems are not the sole responsible.  In many firms the knowledge of the costs of the information system is partial:  only the expenditure on data processing is measured and not those of the business units, which represent between the quarter and one third of the data-processing cost.  Certain firms are let surprise:  their data-processing budget chokes under the cumulative costs of maintenance and the cost of the network of PC

The attention of the managers often concentrates on the projects  and not on the operation  of the firm around its information system.  It is like if, in a city, one were interested in the building operations and not in the life, work and transport of the inhabitants!  That causes inflation.  A project is all the more visible if it is expensive and not if it is useful.  I saw people paying 10 million for an application which could have been produced for 500 000 F using a groupware product.  500 000 F, it was too modest to appear in the grain of the photograph whereas the 10 million, being visible, lent itself to the discussion. 

The larger one project is, the highest the risk of failure.  According to Standish Group 31 % of the large data-processing projects fail completely.  We would not tolerate such a rate of failure in building or industry:  our control of the information systems misses maturity. 

Profitability : Hard to Evaluate

To measure the profitability of an information system implies to measure the profitability of a language and an organization.  However if everyone knows that a badly organized firm is ineffective, nobody computes the ROR of an organizational effort. 

Let us take some marks however.  No Airline can live without reservation system, revenue management and supervision;  no Telco can live without electronic switching equipment or invoicing system;  no bank, no insurance firm can live without a data-processing management of the accounts of the customers.  In these various firms, the profitability of the information system is infinite. 

Econometricians deny however this profitability: they compute averages whereas they should observe only the most mature firms, those which give examples to the others and which will be copied tomorrow.  The statistics allow to evaluate the profitability of an asset only when its use matured and good practices were spread.

Economics indicates that there exists for each firm an optimal rate of computerization as well as an optimal definition of the applicative portfolio.  The firm can approach these optima by tâtonnement, following some simple rules. 

A Stake for Executives

These rules are summarized in three key words:  relevance, sobriety, coherence. 

Relevance  is the adequacy with the needs of the business units.  It is acquired by listening of the experts and benchmarking similar firms, by the evaluation of applications in operation.  It is an experimental approach, freed of the dogmatism which one sometimes confronts in headquarters. 

Sobriety  is complementary of listening.  If one obeys to the letter the request of the users, the system won’t be efficient. One has to define priorities, to simplify as much as possible.  It is partly because our systems are too heavy that they have so often to be rebuild. 

A majority of information systems are subjected to an entropy which degrades their semantic quality:  it appears clearly when one is confronted with incoherent statistics.  Coherence  is obtained by articulating the applications to the reference frame and by placing the data to the centre of the information system. It is the principle of ERP.

A Crisis of Adaptation

These three criteria illustrate the scope of responsibility of the business units.  Data-processing failures were all  caused by unstable business units, which were unable to define their priorities, crossed by childish conflicts.  They missed professionalism. 

Technology gives us today many opportunities, provided that the firm respects the competence of the data processing specialists and is able to listen to them.  Modelling of the information system is delicate, but methods and competences exist.  The largest obstacle is elsewhere. 

Class struggle was displaced in our firms by castes struggle. Professional corporatism and border fights often take the best over cooperation.  This is not a fate.  Our firms go through a crisis of adaptation, of assimilation and understanding of the stakes of information system.  Our conference aims to clarify these stakes and to trace the way towards the resolution of these difficulties.