Wattal, Sunil; Mudambi, Susan; Burtch, Gordon; Schuff, David (David Michael); Kumar, Subodha (Temple University. Libraries, 2020)
      This research examines the relationship between online and offline sales in an omnichannel sales environment centered around the selling of parking spaces. This dissertation consists of a pilot study followed by an expanded study. The studies delved into the effects of e-commerce additions upon traditional brick and mortar revenue channels. The parking industry was selected as the backdrop, given the high degree of current relevant concerns in this space around internet sales-related cannibalization concerns against physical stores, and given the author’s substantial access to relevant research data. Data was collected for an intervention group and a control group from a leading parking management firm with thousands of parking garages located across North America. In the pilot study, and using panel data gathered from the firm, we sought to examine the effects of an online intervention to existing aggregate revenues through the implementation of a new 3rd party e-commerce sales channel (Parkwhiz.com). Under the intervention group, data was collected on 1.7 million revenue transactions over an approximately two-year period, from 2016 through 2018 across 15 parking properties in New York City. The control group consisted of 493,950 revenue transaction entries, also spanning a roughly two-year period from 2016 through 2018, and across 28 different parking properties throughout the City of New York. A fixed-effect model was used for analyzing the data, which came with unforeseen challenges of balance, outlier concerns, and sample size. Ultimately, insignificant results were observed, but these were attributed mainly to difficulties in data structure and sample size (N = 386) of daily revenue observations. Despite those challenges, individual summary statistics showed potential strength in the primary hypothesis, and this motivated further examination. In the expanded study, an adapted approach from the pilot was used to correct for a majority of its shortcomings in the data structure, sample size, balance, and modeling. Further, several moderating components were incorporated to test practice relevant relationships between revenue, the competitive landscape, and online search sessions. Using the same primary hypothesis from the pilot study, the expanded work provided for 15 intervention group properties and 15 control group properties in New York City, with a balanced dataset of 90 days pre-intervention and 90 days post-intervention for each property examined in the year 2018 or 2019. The original hypothesis (H1A) evaluates whether or not an online intervention increases total revenue at a given location. Additional hypotheses (H1B) evaluating whether offline revenue sources are affected by the online intervention, (H2) moderation of revenue by the volume of competition using the same online channel, (H3) moderation of revenue by the volume of parking locations in the marketplace (zip code area) regardless of selling online or offline, and (H4) revenue predicted by the volume of online searches for parking occurring in the marketplace. Sample sizes ranged from N = 3,491 to 5,098 across our various regression models. Our overall H1 outcomes (across four different regression models) showed strong statistical significance with p-values less than 0.001, and moderate R^2 scores between 37%-47% for the online ParkWhiz intervention. Online intervention increases revenue per parking space in the range of $1.171 to $1.196 in the experiment. The results provide support for the proposition that adding an online sales channel to an existing body of physical parking facilities is additive, non-cannibalistic and overall productive for the business. Our H2 and H3 study outcomes were inconclusive, as the moderators were not significant. The tests of the moderating effects in H2 and H3 provided no practical results, other than perhaps anecdotal perception to supplement the other findings. The testing of H4 did show significance in the importance of the assessment of online search demand in a given zip code as an amplifier of the effect of online intervention on parking revenue. Search volume is positively related to a change in the net new revenues. Overall, the analysis generated learnings valuable for future researchers to expand upon through better data gathering, statistical models, and analysis. In totality, the desired contribution of this body of research is to provide today’s brick and mortar business manager with strategic insights into the conditions needed to make healthy e-commerce decisions, based on observable market conditions, in an omnichannel environment that combines online and offline models for maximum aggregate revenue growth. Avoidance or minimization of cannibalization between existing channels and new channels can ensure success. Our work demonstrates several critical aspects of the phenomena of successful online and offline channel cohabitation with practical conclusions for the strategic decision-maker to use in reaching that equilibrium, and leaves a discernible path for future researchers to supplement our efforts with additional moderating variables. Keywords: Omnichannel, cannibalization; externality; brick and mortar; platform; two-sided marketplace; e-commerce; parking; retail; online/offline; distribution; multi-channel; cross-channel; offline-to-online service platform, channel addition, mobile apps.