Research Policy 32 (2003) 533–551 Intrafirm diffusion of new technologies: an empirical application Lucio Fuentelsaz, Jaime Gomez ∗ , Yolanda Polo Departamento de Econom´ ıa y Dirección de Empresas, Universidad de Zaragoza, Gran V´ ıa 2, 50005 Zaragoza, Spain Received 18 May 2001; received in revised form 20 July 2001; accepted 8 May 2002 Abstract Research on the diffusion of new technologies has centred on the study of the interfirm rate of diffusion, paying much less attention to intrafirm aspects. This paper attempts to overcome this gap in the literature by analysing the factors that influence the speed with which a new technology, the ATM, is fully adopted. The data over which the hypotheses are tested belongs to the Spanish savings banks market. The results show that the rate of intrafirm diffusion is explained by innovation, firm and market characteristics. In testing our hypotheses we make use of both traditional methods and survival analysis techniques. © 2002 Elsevier Science B.V. All rights reserved. Keywords: New technology; Intrafirm rates of diffusion; ATM; Savings banks; Survival analysis 1. Introduction Technological progress fundamentally depends on the adoption of new technologies. However, firms nei- ther immediately nor fully adopt them. From a mi- croeconomic perspective, two questions are central to the diffusion debate: (1) What factors influence the adoption of new technologies by the firms operating in a market? (2) What are the factors that influence the speed with which firms tend to fully employ the new technology once they have adopted it? Whereas the first question refers to the interfirm rate of diffusion of new technologies and has been a frequent topic of re- search, much less attention has been paid to the study of the intrafirm rates of diffusion. This is surprising, given that the study of the extent of use of a technology within a firm is at least as important as the study of the number of users of that technology (Stoneman, 1981). ∗ Corresponding author. Tel.: +34-976-761-000; fax: +34-976-761-767. E-mail addresses: lfuente@posta.unizar.es (L. Fuentelsaz), jgvil- las@posta.unizar.es (J. Gomez), ypolo@posta.unizar.es (Y. Polo). The importance of covering this gap in the literature should be emphasised. Given that the great majority of the innovations are divisible, the initial adoption deci- sion constitutes just the first step in a more complex and longer diffusion process by which the potential beneficial effects of a new technology are fully realised to the benefit of society. As Stoneman (1981) points out, this is especially important in a world populated by large firms, in which the initial adoption decision only supposes a small percentage in aggregate terms. This is, in fact, the case in most of the latest techno- logic advances (microcomputers, mobiles, automated teller machines (ATM) or teleprocess terminals) asso- ciated with information processing needs. Consequently, the objective of this paper is to pro- vide evidence on the factors affecting the speed at which new technologies diffuse within firms. In or- der to test the predictions suggested by the theory, we examine the characteristics of the diffusion of a new product and process technology, the ATM, on a sample of Spanish savings banks. This technology has been the subject of several studies (Hannan and McDowell, 1984a,b, 1986; Ingham and Thompson, 1993; Maudos, 0048-7333/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII:S0048-7333(02)00081-1