Contents lists available at ScienceDirect Transportation Research Part A journal homepage: www.elsevier.com/locate/tra A diffusion model for estimating adoption patterns of a one-way carsharing system in its initial years Cen Zhang a, ⁎ , Jan-Dirk Schmöcker a, ⁎ , Masahiro Kuwahara b , Toshiyuki Nakamura c , Nobuhiro Uno d a Department of Urban Management, Kyoto University, Japan b Toyota Motor Corporation, Japan c Institute of Innovation for Future Society, Nagoya University, Japan d Department of Earth Resource Sciences, Kyoto University, Japan ARTICLE INFO Keywords: Adoption Diffusion models Carsharing Social influences Spatial heterogeneity Demand forecasting Short-time intervals ABSTRACT Oneway carsharing service operators must assess the importance of each station relative to overall demand if resources are limited. We propose a variation of an innovation diffusion model designed to estimate new “hesitant” and “fast” adopters for different stations of a one-way car- sharing system to understand system adoption dynamics better over time and to derive policy implications. We forecast the number of monthly new adopters and potential market of stations considering their synergistic effects. We further reflect the spatially diverse adoption dynamics during the initial years of a carsharing service. Stations are classified into four groups based on their location and demand pattern. The models are estimated using data from the Ha:mo RIDE carsharing system in Toyota, Japan. We observe two peaks in the new user curve that our model can explain. We propose that the initial peak is caused by information diffusion, whereas the later peak is due to market saturation. Policy relevant implications are that we observe a low degree of follower effect and that new stations in strategic locations are essential for continued demand growth. More specifically, we suggest that carsharing stations in residential areas experience demand stagnation fast, whereas continued demand growth can be expected and quantified for carsharing stations around transit hubs and public facilities. Therefore carsharing operators need to be aware that assessing the importance of a station for the overall system requires time as well as consideration of synergy effect with other service points. 1. Introduction Shared mobility, which includes carsharing, bicycle sharing, micro-transit, and on-demand ride services, has been spreading rapidly in cities around the world. This is especially the case in cities with space constraints and high transport costs. Such devel- opments do not come without problems for service operators. According to Iimedia Research (2018), a renown data analysis or- ganization focusing on the emerging industries in China, the number of free-floating bike-share users in China increased nearly 10 times between 2016 and 2018, reaching 235 million. This has been achieved by extending free-floating bike-share services to nearly 300 cities and supplying more than 30 million bicycles. However, the ratio of active bicycles per month has dropped below 50%, meaning that most of the resources are wasted. Furthermore, bicycles that are rarely used can often be found broken. This means that https://doi.org/10.1016/j.tra.2020.03.027 Received 21 May 2019; Received in revised form 17 March 2020; Accepted 23 March 2020 ⁎ Corresponding authors. E-mail addresses: zhang.cen@trans.kuciv.kyoto-u.ac.jp (C. Zhang), schmoecker@trans.kuciv.kyoto-u.ac.jp (J.-D. Schmöcker). Transportation Research Part A 136 (2020) 135–150 Available online 10 April 2020 0965-8564/ © 2020 Elsevier Ltd. All rights reserved. T