ATTRIBUTION MODELING AND MARKETING RESOURCE ALLOCATION IN AN ONLINE ENVIRONMENT
MetadataShow full item record
This dissertation contains one conceptual framework and two essays on the attribution modeling and marketing resource allocation in digital marketing. Chapter II presents the conceptual framework for attribution modeling and hypotheses related to the carryover effects and spillover effects of the information collected during the customer's prior visits through different marketing channels to a firm's website on subsequent visits and purchases. In Chapter III, I propose a method to measure the incremental value of individual marketing channels in an online multi-channel environment. The method includes a three-level measurement model of customers' consideration of online channels, their visits through these channels and subsequent purchase at the firm's website. Based on the analysis of customers' visits and purchases at a hospitality firm's website, I find significant carryover and spillover effects across different marketing channels. According to the estimation results, the relative contributions of each channel are significantly different as compared to the estimates from the widely-used "last-click" metric. A field study was conducted where the firm turned off paid search for a week to validate the ability of the proposed approach in estimating the incremental impact of a channel on conversions. This method can also be applied in targeting customers with different patterns of touches and identifying cases where e-mail retargeting may actually decrease conversion probabilities. Chapter IV analyzes the impact of attribution metric on the overall effectiveness of keyword investments in search campaigns. Different attribution metrics assign different conversion credits to search keywords clicked through the consumers' purchase journey, and the attribution-based credits affect the advertiser's future bidding and budget allocation for keywords, and in turn affect the overall return-on-investment (ROI) of future search campaigns. Using a six-month panel data of 476 keywords from an online jewelry retailer, I empirically model the relationship among the advertiser's bidding decision, the search engine's ranking decision, and the click-through rate and conversion rate, and analyze the impact of the attribution metric on the overall ROI of search campaigns. The focal advertiser changed the attribution metric from last-click to first-click half-way through the data window. This allows me to estimate the impact of the two attribution metrics on budget allocation, which in turn influences the realized ROI under different attribution regimes. Given the mix of the keywords bid by the advertiser, the results show that first-click leads to lower overall revenues and this impact is stronger for the more specific keywords. The policy simulation shows that the advertiser would be able to improve their overall revenue by more than 5% by appropriately changing the attribution metric for individual keywords to account for their actual contribution.