We investigate the effect of offline social interactions on online shopping demand and the moderating role of online channel preference in this offline-online relationship. To be specific, we intend to obtain empirical evidence by answering the following questions. First, do offline social interactions affect online demand? Second, to what degree do the active versus passive kinds of offline social interactions have the differential influence on online shopping demand? Third, how does online channel preference affect the effect of offline social interaction on online shopping demand? Drawing on the related literature in the fields of social interactions and Internet retailing, we hypothesize that the active kind of offline social interactions exerts positive influence on online shopping demand whereas the passive kind of offline social interactions has negative effects. We further hypothesize that online channel preference weakens the influence that offline social interactions has for online shopping demand. Both the positive impact of active interactions and the negative impact of passive interactions diminish in determining online shopping demand as online channel preference gets greater. We obtained sales data between January 2008 and April 2010 from a leading Internet retailer that sells baby products in the U.S. The data includes the information of zip codelevel sales and shipping days. We merged this proprietary data with the following three commercial datasets purchased from ESRI (Environmental Systems Research Institute): (1) 2011 Civic Activities Market Potential, (2) 2011 Internet Market Potential, and (3) 2011 Baby Products Market Potential). Each of these datasets includes the information of offline social interactions, online shopping preferences and offline baby product sales, respectively. Finally, as we focus on the zip code-level interplay between offline social interactions and online demand, we control for regional demographics and market condition. As such, we obtained the 2010 Census data and 2009 ACS (American Community Survey) data to account for overall local environments (e.g. population density of children aged less than five years, percentage with college education). Our empirical analyses and hypotheses testing provide the following important findings. First, active offline social interactions have positive effects on online shopping demand. This indicates that active social interactions reflect information exchange among long ties, and this informational influence in turn reduces any risk and uncertainty associated with online shopping. Second, passive offline social interactions have negative effects on online shopping demand. This suggests that passive social interactions take place among local ties and generate normative influence to conform to the expectations of others about shopping behavior, making online shopping as a new channel less attractive there. Third, online channel preference is significantly positive on online shopping demand, confirming prior studies on the relationship between channel preference and demand (Changchit et al. 2014; Valentini et al. 2011). Fourth, the positive effect that active offline social interactions have for online shopping demand decreases as online channel preference increases. Regions with strong online channel preference are likely to have well-established channel propensity and the informational influence of social interactions in reducing uncertainty becomes weaker. As such, social interactions do not play a role in spreading information about the online marketplace in regions where online channel benefits are well understood (Burt 1992, 2005; Harrigan et al. 2012). Lastly, the negative influence of passive offline social interactions gets smaller as online channel preference gets greater. Online channel preference reflects the locally-determined attractiveness of the online marketplace, and this in turn weakens normative influence to conform to the expectations and shopping behaviors of local ties.

• Jeeyeon Kim(Yonsei University, Republic of Korea)
• Mingyung Kim(UC Berkeley, USA)
• Jeonghye Choi(Yonsei University, Republic of Korea)