Existing theories on the diffusion of innovations fail to sufficiently account for contextual factors such as institutions, infrastructure, and supply-side dynamics. This paper presents a novel framework to analyse technology diffusion from a sociotechnical systems perspective, intended as an analytical tool to identify and assess drivers and barriers to diffusion that could be addressed through policy or business strategy. This framework, referred to as the diffusion innovation system (DIS) approach, is positioned within the innovation systems literature. The framework is applied to two empirical cases of renewable energy technology diffusion in Sweden: solar photovoltaics (PV) and wind power. The cases illustrate how key factors related to institutions, infrastructure, adopters, and supply co-develop over time as the technologies diffuse, hence demonstrating the merits of the framework. As these changes are both a reaction to and a cause of diffusion, the sociotechnical diffusion system develops through positive feedback loops. Although the systems' development is largely conducive of diffusion, some remaining and potential barriers are identified.
The diffusion of innovations – i.e., the process through which a new technology is adopted by an increasing number of actors throughout society – is important for economic, social, and ecological sustainable development. For example, to mitigate climate change, renewable energy technology diffusion is needed. Understanding how different factors shape diffusion patterns is important as this helps policy makers and business leaders facilitate diffusion.
There is a broad and diverse literature on how innovations emerge and diffuse. First, sociotechnical systems approaches to technological transitions (Markard and Truffer, 2008; Weber and Rohracher, 2012) such as the multi-level perspective (e.g. Geels, 2010, Geels, 2002; Schot and Kanger, 2018; Smith et al., 2010) and different innovation system frameworks (e.g. Binz and Truffer, 2017; Carlsson and Stankiewicz, 1991; Edquist, 1997) emphasise factors beyond the diffusing technology itself and its adopters, including dynamic and complex relationships between institutions and actor networks. These approaches traditionally treat technology development, production, and diffusion as occurring together in one and the same system, typically not emphasising the diffusion part. While these approaches suit immature technologies struggling to diffuse beyond their earliest applications, they are not appropriate to analyse the diffusion of more mature technologies (Mignon and Bergek, 2016). For example, the technological innovation systems approach has been widely used as a hands-on analytical tool for informing policy on how to support e.g. renewable energy technologies and electric vehicles in early stages of these technologies' development. However, several sustainable technologies are now mature and (ready to) diffuse beyond their initial markets. Hence – as stated in recent calls for research in the sociotechnical transitions literature (Köhler et al., 2019; Mignon and Bergek, 2016; Palm, 2017a) – there is a need for a diffusion-focused system-oriented analytical tool.
Synthesis and conclusions
The two case studies illustrate that the DIS framework is appropriate for analysing the diffusion of innovations from a sociotechnical systems perspective. The cases reveal that, as the technologies have diffused, much development has occurred through the DIS framework's four key processes. This development has been crucial to facilitate diffusion, although some barriers to diffusion remain. The processes both induced and responded to diffusion, and they developed symbiotically fuelling one another. Hence, positive feedback mechanisms were important for the system's development. Overall, the national system boundary made sense for the two cases as the processes' development was – besides the import of standardised artifacts – largely determined by factors within the system. Although the system boundaries were by no means perfect, they proved good enough to serve as analytical constructs.