Successful implementation of new technologies is necessary for survival in today's increasingly competitive banking environment. The banking sector is now the UK's largest investor in new technology. As a result of de-regulation and new technological opportunities, the dominant market position of the traditional banks is under threat from new market entrants. Such organisations are harnessing the latest technology to develop their services, and are not constrained by the costs of running national branch networks and integrating a diverse historical legacy of incompatible computer systems. This thesis addresses the question of why the massive investment by UK retail banks in information technology is not being translated into significant productivity gains. The issue is referred to in the literature as the IT productivity paradox. It will be argued that the importance of the banks to the economy as a whole, and the position of the industry as the UK's largest investor in new technology, make it a particularly suitable arena in which to study this phenomenon. Six empirical case studies of recent new technology projects are analysed by the development of grounded theory. The findings indicate that dysfunctional organisational structures and cultures, together with poor management of expertise, combine to sabotage change and constrain the potential of new technology projects. Most significantly, a lack of organisational learning is contributing to the paradox. Even successful projects had limited impact because the lessons learned were not disseminated throughout the organisation. The study concludes that the full potential offered by information technology will continue to elude the banks until their apparently complacent attitude towards organisational change is addressed.

Les travaux autour de l’instrumentation gestionnaire sont le point d’entrée de cette recherche (Moisdon et al., 1997 ; Hatchuel et Weil, 1992 ; David, 1996, 1998). Nous nous intéressons à la coévolution d’une pratique de gestion, le knowledge management (KM), et d’une organisation. En résumant nos questions de recherche par la formule : « comment peuton faire du KM ? », nous interrogeons à la fois les différentes approches possibles de la gestion des connaissances, mais aussi les raisons qui conduisent une organisation à adopter le KM. Pour ce faire, nous menons une analyse multi-niveaux qui tient compte du contexte institutionnel, organisationnel et interindividuel. Dans ce cadre, nous mobilisons plusieurs littératures - institutionnalisme, outils de gestion, modes de consommation, théories managériales des organisations multinationales, régulation sociale - afin de parvenir à une compréhension globale des dynamiques d’un dispositif de gestion des connaissances au sein d’une organisation. Les dynamiques du dispositif de gestion sont étudiées à travers la mise en intrigue (Ricoeur, 1983) de la gestion des connaissances dans la Branche Ciment du groupe Lafarge. L’intrigue - pourquoi le KM ? - s’articule autour des enjeux stratégiques de l’industrie cimentière et de l’organisation de la fonction d’expertise technique. .Elle est le prélude aux différents récits qui content l’histoire du KM dans la Branche Ciment selon différentes perspectives spatiales et temporelles. A l’issue de cette recherche, il s’avère que le rôle principal du dispositif de gestion des connaissances est celui d’accroître la légitimité de l’entreprise vis-à-vis de son environnement externe et interne. Le dispositif KM est un outil au service d’un centre fort, qui se veut détenteur des connaissances. Nous avons documenté le basculement du dispositif de gestion fondé sur une logique de personnalisation, de gestion des compétences et d’autonomie vers une logique de codification, de gestion de l’information et de prescription. Le dispositif, figé dans une régulation de contrôle, se heurte à la régulation autonome des membres de l’organisation qui s’exprime à travers la résistance à la codification. Le conflit entre ces deux régulations entraîne une crise qui se traduit notamment par le mal-être des experts. Cependant le découplage soigneusement maintenu entre le dispositif de gestion des connaissances et l’activité réelle des opérations permet le maintien du statuquo. Au final, le cas de la Branche Ciment illustre un cas de domination stratégique. En l’absence de régulation conjointe, c’est le découplage entre discours et pratique qui assure la survie du modèle d’organisation. / This research takes the literature on management tools (Moisdon et al., 1997 ; Hatchuel et Weil, 1992 ; David, 1996, 1998) as a starting point. I focus on the co-evolution of a management practice in-use (Jarzabkowski, 2004), here Knowledge Management (KM) and an organisation. My research deals with the question « How can we do KM ? ». This refers to the different approaches for managing knowledge as well as to the different reasons that lead an organisation to adopt KM practices and tools. A multi-level analysis, combining the institutional, organisational and inter-individual contexts, is carried out. This framework calls upon various theories: institutionalism, management tools, managerial theories of the multinational corporation, social regulation (Reynaud, 1989), thus broadening our understanding of the dynamics of a management practice in-use within an organisation. The concept of emplotment (Ricoeur, 1983) is used to study the dynamics of KM within the Cement Division of the Lafarge group. The main plot - why KM? - brings forward the strategic context of the cement industry and the organisation of the technical expertise function. This plot is the starting-point for the different stories of KM in the Cement Division, which are told from different perspectives in time and space. The main results of this research show that Knowledge Management is mainly used to increase the firm’s legitimacy both towards external and internal stakeholders. KM is a tool in the hands of strong headquarters that claim to be the ultimate knowledge owner. Thus, I related the switch of KM from a perspective based on personalisation, competencies management and autonomy to a perspective focusing on codification, information management and prescription. Resistance to codification highlights the conflict between KM, set in a control mode (régulation de contrôle) and the employee’s will for autonomy (régulation autonome). The tension between these two types of regulation leads to a crisis as can be seen through the growing sense of unease among the experts. In the end however, such tensions do not interfere with the daily run of operations as KM is decoupled from ongoing activity. Thus, the story of the Cement Division is a case of strategic domination. In the absence of a joint regulation (régulation conjointe), decoupling discourse from practice allows the survival of the organizational model.