Essays in digital advertising and offline retailing

Abstract

This dissertation consists of two essays, one on estimating and optimizing online multiple-channel advertisement and one on identifying the pandemic's persistent impact on consumers' behaviour in offline retailing. The first chapter explores budget allocation strategies for a multichannel ad campaign. A marketing agency strives to maximize the total conversions by dynamically adjusting budget allocation over marketing channels. A salient feature of the problem is the interplay of spillover and carryover effects; namely, customers are exposed to ads through multiple channels, thus ads from one channel affect the effectiveness of the subsequent ads from other channels. We derive a fluid approximation to consumer dynamics across multiple channels, which refers to characterizing structural properties of optimal dynamic budget allocation policies that internalize the cross-channel interactions. To enable practical implementation, we propose a static budget allocation policy that is both tractable in practice and near-optimal for long campaigns. Our theoretical results provide normative guidance for budget allocation in multichannel ad campaigns. We illustrate the efficacy of our proposed method through a numerical empirical study based on data from an online multichannel ad campaign. The second chapter presents novel findings on a sustained increase in customers' basket size in the convenience store industry due to the COVID-19 pandemic. During the pandemic, customers faced policy restrictions, notably enlarging their basket sizes while reducing purchase frequencies for offline shopping. Unlike other offline sectors where basket sizes returned to pre-pandemic levels after restrictions were lifted, the larger basket size phenomenon endured in the convenience store industry. We utilized county-day level data to examine this phenomenon and employed a difference-in-differences analysis, using the year before the pandemic as a control group. Our analysis revealed a significant 17% increase in basket size, as measured by gross spending per order, even 19 months after the policy shock.

Date of Award	2024
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology
Supervisor	Ying Ju CHEN (Supervisor) & Dongwook SHIN (Supervisor)