Building an innovation discontinuance model : the case of twitter




Ng, Yee Man

Journal Title

Journal ISSN

Volume Title



This dissertation seeks to extend Everett Rogers’s Diffusion of Innovations theory by examining social media users’ post-adoption behavior. Despite the rapid growth of social networking sites (SNSs), the rate of user discontinuance is staggering. Keeping users active and engaged has always been a crucial issue for SNSs. Prior diffusion research has largely focused on innovation adoption, whereas innovation discontinuance is overlooked. However, innovation discontinuance is a vital facet of the diffusion process. In the real world, only a few innovations become institutionalized while most end up being fads that most users discontinue quickly. While early studies approached discontinuance as a one-time, complete abandonment of an innovation, this study extends the concept by examining two types of discontinuance: intermittent and permanent. Intermittent discontinuers are users who leave an innovation for a break but resume the use at a later time; permanent discontinuers are those who have no intentions to return. This study takes a mixed-methods approach—combining a user survey with computational analyses of “big data” drawn from Twitter—to explore the differences between intermittent and permanent discontinuers in three dimensions: (1) their distinctive characteristics (demographic, behavioral, and psychographic), (2) reasons for discontinuance, and (3) decision processes. The concept of intermittent discontinuance leads to the development of a new post-adoption decision-making model, which accounts for discontinuers’ planned and unplanned readoption behavior. This cyclical, multi-stage model also provides a systematic framework to compare the behavior and cognitive reasoning between intermittent and permanent discontinuers at each phase of the post-adoption cycle. While prior studies employed both qualitative and quantitative research methods to examine discontinuance, few came up with clear and reliable ways to measure the timeframe of discontinuance and users’ reasons for discontinuance. To address the arbitrariness of determining what length of inactivity constitutes intermittent and permanent discontinuance, this study introduces a mathematical approach based on an innovation’s life cycle and its user base. To examine users’ reasons for discontinuance, this study refines and expands Rogers and Shoemaker’s replacement-disenchantment typology—by factors and by discontinuance typologies. While Rogers conceptualized the innovation-diffusion process as an uncertainty reduction process, this study suggests that post-adoption decision-making process is a disturbance-coping mechanism—a temporal settlement of the constant interplay between an innovation’s utilitarian performance and social media exhaustion. Intermittent discontinuance usually occurs due to information overloads. Permanent discontinuance tends to occur due to perceived innovation shortcomings and innovation replacement. This dissertation provides theoretical insights into the temporal instability of an innovation, and why and how an innovation is discarded or discredited. The findings contribute to an adequate comprehension of the entire innovation diffusion process, which also helps SNS providers develop tailor-made retention solutions to re-engage SNS users.


LCSH Subject Headings