Integrated Machine Learning Models for Text and Consumer Journey Analysis
Understanding the consumer purchase journey is crucial yet challenging due to prolonged purchase cycles, unobserved decision-making stages, and the impending loss of third-party cookie-level data. This research integrates a Hidden Markov Model (HMM), topic models, and autoencoders to address the extensive touchpoints along consumers’ purchase journeys. The HMM connects fragmented data across various granularity levels, uncovering consumers’ latent purchase states and capturing the sequences and dynamics of their touchpoints. Meanwhile, the topic models and autoencoders manage the diverse nature of touchpoint data from multiple sources, reducing data dimensionality while preserving meaningful insights.
We demonstrate an application of machine learning methods to analyze the consumer purchase journey, particularly in the context of privacy-related challenges, the loss of third-party cookie-level data. By integrating textual analysis and deep learning with firm-level decision-making needs, this research offers a comprehensive framework for combining high-dimensional and fragmented data. It also simulates predictive accuracy loss in marketing analytics without third-party cookies, effectively linking granular user-level data with aggregated data to support actionable marketing insights.
Alice Li
Professor Alice Li joined the Fisher College of Business at The Ohio State University in 2017, after serving on the faculty at Indiana University from 2014 to 2017. She earned her Ph.D. in Marketing from the University of Maryland – College Park in 2014.
Her research focuses on the consumer purchase journey, emphasizing marketing effectiveness through marketing mix models (MMM) and multi-touch attribution (MTA). Recently, she has concentrated on helping firms address challenges related to fragmented data, privacy regulations, and predictive analytics in marketing. Her work includes: (1) measuring the consumer purchase journey with MMM and MTA, (2) initiating the journey through acquisition strategies such as sampling, free trials, and freemium models, and (3) advising firms on navigating disruptions in the consumer journey, such as radical innovations. She applies Bayesian statistics, econometrics, machine learning, and causal inference to real-world data across sectors like hospitality, software, banking, and publishing.
Professor Li’s research has earned over 3,300 Google Scholar citations and 10,000 SSRN downloads. She is a recipient of the MSI Young Scholar Award and a two-time finalist for the Paul Green Award. Additionally, she has received the IJRM Best Article Award, the Adobe Digital Marketing Research Award, and several research fellowships and grants. Her work is published in leading journals, including Marketing Science, Journal of Marketing Research, and Production and Operations Management.
Alice Li | Fisher College of Business