Essay on Customer Engagement Analytics in Social Media Using Deep Learning

Brands are increasingly using social media platforms such as Facebook and Twitter to promote their services and products and engage with their customers. With the increasing prevalence of various social media platforms, intensive online marketing competition has emerged among companies. How to target potential customers in social media effectively and efficiently is an important yet unsolved problem. In the meantime, some newly emerged formats of advertisement have become indispensable components of online marketing, such as influencer marketing where brands collaborate with social media influencers to broadcast products or services. While the usefulness of leveraging influencer endorsements from the standpoint of a brand or corporation has been well examined, a crucial but unanswered topic is how commercial endorsements affect the influencers themselves. In this dissertation, I explore the algorithms and strategies that benefit brands and content providers in the competitive environment on social media. Specifically, from an algorithmic perspective, I develop novel Artificial Intelligence frameworks with interpretability grounded in behavioral and social theories to target potential customers on social media, which in turn prompt brand profitability and market performance. In addition, I combine econometrics with deep learning, text mining, and computer vision techniques, to understand the mechanism of online viewer behavior for influencer marketing. This study contributes to our understanding of how to engage customers for online marketing. The outcome provides stakeholders with a strong forecast and a comprehensive understanding of customer engagement mechanisms. This dissertation also brings practical information in the competitive climate of online marketing.