Traditionally, research works related to quality in telecommunication services have consisted of a series of analysis over a set of critical-to-quality technical parameters and a resulting enhancement proposal. However, in this Phd. Thesis we aim to face the quality issue from an overall point of view: Not only analysing the objective performance of those agents involved in service provision but also by trying to assess which is their effect in final users' satisfaction, since this must be the final goal of this kind of systems.
We have carried out a State of The Art focused on four different areas: technical QoS, subjective quality assessment models, corporative quality management systems and integrating approaches. Based on this work we have identified several deficiencies that will determine the characteristics that the model should show. Among these deficiencies pointed out, the lack of a methodological and quantitative analysis of the real impact of the technical performance over users' satisfaction stands up above all. There is a clear need for a model that allows to represent different quality aspects with an integrating approach. This new model should provide different user profiles (both technical and non technical ones, providers, end-users and regulation agencies) with useful information.
Our proposal comprises both a theoretical model and an application methodology that eases up its use by other researchers. We propose a matrix model representation. One of the sides illustrates those services worth-analysing from an end users' point of view. In this vertical axis each service is composed by a set of perceptions. These perceptions represent the way users measure their satisfaction for this service. On the other axis, the objective point of view is shown. This axis is composed by both the agents that take part in service provision and their set of capabilities. Finally the matches show the presence of a dependency between both realities.
We have also developed an application methodology, that defines a set of steps to follow in order to build up the model. It is a backward methodology, since it starts from the subjective satisfaction, to finish identifying the performance indicators that cause this satisfaction.
Finally, the diagnosis methodology comprises some tools that have proved to be useful so as to extract information from the model. In order to check the utility of both the proposed model and the methodology we decided to use them to illustrate one of the study cases that motivated in first place our thesis work. This way, we will be able to probe that our proposal let us represent the variation of end-users' satisfaction as a function of a given agent's behaviour as well as identify the responsible for this satisfaction every time.
Author of the PhD:
Fidel Liberal Malaina
University of the Basque Country
fidel.liberal@ehu.es