Marketing Intelligence
1. Effectiveness and Policy Design
How can firms measure the true impact of major marketing decisions—and design better policies in response? This research area examines the effectiveness of key marketing levers, including advertising, pricing, loyalty programs, platform fee structures, and digital interfaces. Using structural econometric models and causal inference methods, we quantify how these policies affect consumer behavior, firm performance, and market outcomes. The goal is not only to evaluate what works, but also to guide the design of more effective and profitable marketing policies.
Sample research topics:
- Coordinating cross-channel advertising strategies across TV and digital media
- Designing loyalty programs to optimize firm revenue and customer value
- How TV advertising affects the distribution of new products in B2B markets
- How consumers respond to prices over product lifecycle
- How delivery fee policies affect purchases, revenue, and operating costs
- How search interface design influences consumer search and preferences
- How recommendation algorithms shape user consumption and contribution
- How persuasive strategies influence sales in livestream e-commerce
2. Platform Data Analytics
What can firms learn when data extends far beyond structured tables into massive streams of text, images, video, and behavioral traces? This research area develops methods to extract decision-relevant intelligence from large-scale platform data. By combining topic modeling, deep learning, and LLMs, we uncover consumer preferences, latent needs, intentions, and risk signals from user-generated content, search queries, clickstream behavior, and reviews/complaints. The broader aim is to transform complex, high-dimensional digital data into actionable insights for marketing and platform strategy.
Sample research topics:
- Estimating consumer preferences from online search queries and clickstream data
- Understanding users' multifaceted preferences on digital platforms
- Screening emerging risk signals from large volumes of user-generated content
- Designing localized assortment strategies using large-scale behavioral data
- Building agentic AI systems for automated marketing video analytics
3. Branding & Global Strategy
In the digital era, brands are shaped by continuous streams of content, interaction, and cultural interpretation—not just by slogans or campaign messages. This research area investigates how firms can measure, manage, and scale brand meaning in dynamic and global environments. We develop multimodal AI frameworks to quantify brand equity from social media and digital content, generate on-brand creative assets, and support data-driven strategies for cross-cultural brand localization. Our goal is to make brand strategy more observable, measurable, and actionable in the age of digital platforms and global markets.
Sample research topics:
- Understanding how digital content ecosystems reshape brand perception and consumer engagement
- Quantifying brand equity in real time on social media e-commerce platforms
- Generating visual advertising assets that communicate brand persona and positioning
- Designing brand strategies for cross-cultural adaptation and global market expansion
4. Generative AI for Marketing Decision Science
We do not only study generative AI as a phenomenon—we build and apply it to solve real business problems. This research area explores how LLMs, diffusion models, GANs, and related generative methods can enhance marketing analysis, experimentation, and decision-making. From synthetic data generation to controllable creative design, we develop AI systems that help firms make faster, more scalable, and more personalized decisions. This stream connects methodological innovation with practical business applications.
Sample research topics:
- Prompt engineering for narrative and content marketing using LLMs
- Training synthetic experts for domain-specific screening and evaluation tasks
- Augmenting panel data for high-dimensional retail decisions using GANs
- Generating controllable visual stimuli for behavioral experiments using diffusion models
- Developing preference-aware diffusion models for personalized ad image generation