Conferences
MSI 2023 Analytics Conference: Technology, New Data Streams, & Marketing Strategy
MSI’s annual industry-leading symposium for analytics leaders, data scientists, and top researchers. Scheduled to take place May 4-5 in Philadelphia, PA at the Wharton School of the University of Pennsylvania.
The conference will report the latest developments in MSI’s Marketing Mix Modeling Initiative.
Your marketing strategy is only as good as your data and analytics. As old data sources disappear, you may have too little—and as new sources emerge, too much. If you can only manage what you can measure, how can you manage all that data and the measurement process itself?
In this conference, you will engage with top academics and business leaders who are at the forefront in applying artificial intelligence and advanced analytics to address critical challenges facing marketing today:
- How can deep learning models analyze unstructured data to identify “interpretable spatial representations” of opportunities for innovation and patents, gaps in skillsets or strategic challenges?
- What is the role of micro and macro-influencers in targeting audiences and diffusing new products or ideas?
- As consumers turn to livestream shopping, how will AI assist dynamic customer interactions?
- How are companies using large language models such as ChatGPT to automate creative content and other marketing activities?
- Can incorporating theory on emotional response make “black box” predictive models more explainable and generalizable?
- What are best practices for measuring and managing customers’ digital experience across touchpoints and over time?
- How should behavioral scientists and data analysts collaborate to improve user experience with new technologies?
Daily programming will take place at the Wharton School of Business:
Thursday, May 4th
Jon M. Huntsman Hall (JMHH)
8th floor
3730 Walnut Steet
Philadelphia PA, 19104
Friday, May 5th
Jon M. Huntsman Hall (JMHH)
Room 265
3730 Walnut Steet
Philadelphia PA, 19104
The MSI 2023 Analytics Conference is included in MSI Corporate Membership.
Thursday, May 4, 2023
10:30 – 11:00 a.m. Registration
11:00 – 11:15 a.m. Welcome and Introductions
John Lynch, University of Colorado and Executive Director, Marketing Science Institute
11:15 – 11:45 a.m. Driving Participant Engagement & Outcomes
Kristin Federico, Head of Participant Marketing & Communications, Vanguard & Gregory Miller II, Senior Manager, Client Experience Analytics, Vanguard
During our discussion, we’ll be highlighting how Vanguard leverages the rich data and analytics we have on our plan participants to offer up hyper-personalized experiences through Artificial Intelligence. Our use of behavioral science within our participant experiences, combined with our data and machine learning models drives better financial outcomes for our participants and continues to be a delighter for clients and differentiator in our industry.
11:45 a.m. – 12:15 p.m. Livestream Shopping and Dynamic Customer Interactions
Xiao Liu, Associate Professor of Marketing, New York University
Nowadays consumers use digital devices wherever they go, so marketers can engage consumers with frequent and dynamic marketing interventions. Xiao Liu will present a dynamic personalized pricing strategy (or dynamic coupon targeting strategy) based on a state-of-the-art artificial intelligence algorithm, deep reinforcement learning, which also empowers AlphaGo. A large-scale field experiment conducted in collaboration with the world’s largest live-stream shopping platform demonstrates that the proposed dynamic personalized pricing strategy outperforms all the benchmarks.
12:15 – 1:15 p.m. Lunch with Speakers
1:15 – 1:45 p.m. Applying “Explainable” AI: Using Theory to Understand AI Emotion Models
Hortense Fong, Assistant Professor of Marketing, Columbia University
Images, music, speech, and text often elicit emotional responses in customers. Many machine learning models have been proposed to predict these emotional responses for subsequent downstream tasks. In particular, deep learning models often prove to have good predictive accuracy. However, these models are often also viewed as opaque black boxes, creating concern as to their generalizability. Incorporating theory about emotions into AI can help alleviate these concerns by making such models more explainable while maintaining good predictive accuracy. Incorporating theory about emotions into AI can help alleviate these concerns by making such models more explainable while maintaining good predictive accuracy.
1:45 – 2:30 p.m. MSI’s Marketing Mix Modeling Initiative Progress
Elea Feit, Associate Professor of Marketing, Drexel University & Alice Li, Associate Professor of Marketing, The Ohio State University
Increasing privacy concerns, an increasing array of MMM vendor solutions, and multidimensional data that can vary in granularity, time, markets, and other characteristics, have made it difficult for industry partners to navigate best practices in assessing marketing spend. The Marketing Science Institute will report on Phase 1 of the MSI MMM Initiative, identifying MMM, outlining industry use cases in scorecards, forecasting, and optimization, and identifying MMM design considerations and opportunities for innovation. An introduction and discussion about how member companies can get more deeply involved in the MSI MMM Initiative Phases 2 and 3, focused on data and validation, will follow.
2:30 – 2:45 p.m. Networking Break
2:45 – 3:45 p.m. “Responsible AI” and Data Ethics
Miriam Vogel, President and CEO of EqualAI & JoAnn Stonier, Chief Data Officer, Mastercard; Moderator Mary Purk, Executive Director, AI and Analytics for Business, The Wharton School
With the rapid development of AI technology, it is increasingly important for both business leaders and academics to address the ethical implications and risks associated with the use of AI in the marketplace.
- What role will “Responsible AI” and Data Ethics play in shaping our marketplaces and society as a whole?
- What should leaders be worried about most when it comes to AI algorithms and Data Ethics?
- What are best practices for collecting and managing data access that minimize bias, prioritize privacy and mitigate potential risks for enterprises?
- How should we design and implement AI systems that prioritize these ethical considerations and mitigate potential risks?
This discussion will provide valuable insights and practical advice for company leaders and academics who are navigating the complex landscape of AI Ethics and Responsible AI.
3:45 – 4:15 p.m. AI and the Future of Online Retailing Shopping
Nitzan Mekel-Bobrov, PhD., Chief AI Officer, eBay
Nitzan will discuss eBay’s use of generative AI to enrich the online shopping experience to make online shopping less transactional and more experiential, more like shopping at a physical store, to create live streamed events, and to deliver personalized experiences to create higher engagement and more direct interactions with buyers.
4:15 – 4:40 p.m. Wrap Up
Raghu Iyengar, Conference Co-chair, Professor of Marketing, University of Pennsylvania
4:40 – 4:50 p.m. Alden G. Clayton Doctoral Dissertation Proposal Competition – 2022 Award Ceremony
John Lynch, University of Colorado and Executive Director, Marketing Science Institute & Sherry Pincus, Managing Director, Marketing Science Institute
4:50 – 5:00 p.m. Marketing Mix Modeling Initiative Future
Scott McDonald, President & CEO, Advertising Research Foundation (ARF)
Learn how to engage with the next steps in the MMM Initiative through the development of scorecards, benchmarking, and validation.
5:00 – 7:00 p.m. Networking Reception
This welcome reception is a chance to meet the speakers and other attendees. The AGC winner and honorable mentions will present their research through poster sessions.
Balancing Privacy and Personalization
Malika Korganbekova, Kellogg School of Management, Northwestern University, AGC 2022 Winner
The project investigates the impact of privacy restrictions on online retail platforms’ ability to personalize user experience and finds that in the absence of privacy restrictions, personalization leads to 10% lower product returns and higher consumer welfare. However, in the presence of privacy restrictions, distorted signals of consumer preferences result in substantial welfare losses, which could be mitigated by segmenting consumers based on interests following Google’s Privacy Sandbox solutions.
The Challenges of Deploying an Algorithmic Pricing Tool: Evidence from Airbnb
Mohsen Foroughifar, University of Toronto, AGC 2022 Honorable Mention
In this paper, we study the deployment of an algorithmic pricing tool, Smart Pricing (SP), on Airbnb’s platform. Our findings indicate that the hosts who stand to gain the most from the adoption of SP are the least likely to adopt it due to a disparity between the perceived and actual benefits of the algorithm. Through our counterfactual analyses, we demonstrate that the platform can improve the results of the deployment for both the hosts and itself by educating hosts about how the algorithm works and by integrating cost estimates obtained from a structural econometric model with the existing machine learning algorithm.
Understanding Product Trends
Maren Hoff, Columbia University, AGC 2022 Honorable Mention
Marketers need to understand the origins and evolutions of styles to differentiate short-term fads form long-term and recurring trends that remain relevant to consumers and promise market success. Using a multimethod approach that integrates qualitative methods such as interviews and surveys and quantitative methods such as secondary data analyses and algorithmic frameworks, my paper aims at examining the evolution of product trends over time.
Health Insurance and the Dynamics of Patient Decision Making
Jong Yeob Kim, NYU Stern, AGC 2022 Honorable Mention
In the dissertation, I explore the impact of government insurance expansion for elderly care in South Korea. This article exploits variations in eligibility age thresholds and coverage amounts in the South Korean dental market to examine health care utilization, strategic delays, and the dynamics of treatment choices by the elderly patients.
Friday, May 5, 2023
8:00 – 9:00 a.m. Networking Breakfast
9:00 – 9:05 a.m. ARF WIDE – the Workforce Initiative for Diversity and Excellence
Scott McDonald, President & CEO, Advertising Research Foundation (ARF)
9:05 – 9:15 a.m. Opening Remarks
Eric T Bradlow, Conference Co-chair, Professor of Marketing, University of Pennsylvania
9:15 – 9:45 a.m. Using Simulated Data to Validate Marketing Mix Models
Jessica Nguyen, Quantitative Researcher, Meta
Recently, there has been renewed interest in Marketing Mix Modeling (MMM) due to changes in regulations and privacy expectations that impact the way data fuels advertising. However, this increased interest has contributed to questions around how to validate these models. This session will discuss one approach used in the Marketing Science Institute’s Marketing Mix Model Industry Challenge: simulated data. We discuss how to generate simulated data using siMMMulator, an open-source code package, and how to use that data to validate MMMs. Additionally, we consider the wider importance of ground truth data in measurement.
9:45 – 10:15 a.m. Mega or Micro? Influencer Selection Using Follower Elasticity
Ryan Dew, Assistant Professor of Marketing, University of Pennsylvania
In the quickly growing space of influencer marketing, one common criterion companies use to select influencer partners is popularity: while some companies sponsor “mega” influencers with millions of followers, others partner with “micro” influencers, who may only have several thousands of followers, but may also cost less to sponsor. In this research, Ryan Dew and his coauthors develop a framework for navigating this trade-off, and quantifying the returns to influencer popularity. They use this methodology to develop guidelines as to which companies and campaigns may benefit most from mega versus micro influencers.
10:15 – 10:45 a.m. AI at the Front End of Innovation
Jessica Yankell, Director, PBNA Innovation Insights & Strategy, PepsiCo
AI is changing the way we innovate, bringing the consumer and other demand signals into the earliest stages of the innovation process. We will discuss various ways in which AI can bring data, insights, and agility into this process to help us explore and iterate in a more informed way. These include leveraging a wider variety of datapoints and signals, helping to synthesize massive amounts of unstructured data to uncover and predict trends, as well as enhancing our understanding of consumer segments and preferences within the context of prioritized innovation dig sites.
10:45 – 11:00 a.m. Networking Break
11:00 – 11:30 a.m. InnoVAE: Generative AI for Patents, Innovations, and Firms
Dokyun Lee, Associate Professor of Information Systems, Boston University
A lack of interpretability limits the use of common unsupervised machine learning techniques in contexts where they are meant to augment managerial decision-making. DK Lee will present a generative deep learning model based on a Variational AutoEncoder (“InnoVAE”) that converts unstructured patent text into an interpretable spatial representation of innovation (“Innovation Space”) to enable high-resolution exploration at the patent, innovation, and firm level. The same methodology can be applied to various multi-modal business objects such as companies (into strategy space), jobs (into skill space), products (into feature space), and more.
11:30 a.m. – 12:10 p.m. How Behavioral Science Can Help Data Analytics
Stefano Puntoni, Professor of Marketing, University of Pennsylvania & Shiri Melumad, Assistant Professor of Marketing, University of Pennsylvania, & Patrick Moriarty, SVP of Analytics Practice in North America, Kantar
Realizing the full potential of AI and data science to improve business decision making requires analysts to understand how managers will use these resources and how the public will respond. That’s why it’s critical to invite behavioral scientists to join the conversation about analytics and digital technology in general. This discussion will draw on the rich literature on human factors, user experience (UX), and adoption of technology that can guide commercializing user-facing solutions and emerging techniques to leverage new data streams to improve customer experience.
12:10 p.m. Adjourn
John Lynch, University of Colorado and Executive Director, Marketing Science Institute
Recommended Hotels:
- Sheraton Philadelphia: 3549 Chestnut St, Philadelphia, PA 19104, (215) 387-8000
- The Inn at Penn: 3600 Sansom St, Philadelphia, PA 19104, (215) 222-0200
- The Study at University City: 20 S 33rd St, Philadelphia, PA 19104, (215) 387-1400
Presentations
Livestream Shopping and Dynamic Customer Interactions – Xiao Liu, New York University
Applying “Explainable” AI: Using Theory to Understand AI Emotion Models – Hortense Fong, Columbia University
MSI’s Marketing Mix Modeling Initiative Progress – Elea Feit, Drexel University & Alice Li, The Ohio State University
“Responsible AI” and Data Ethics – Miriam Vogel, EqualAI & JoAnn Stonier, Mastercard; Moderator Mary Purk, AI and Analytics for Business, The Wharton School
AI and the Future of Online Retailing Shopping – Nitzan Mekel-Bobrov, eBay
Marketing Mix Modeling Initiative Future – Scott McDonald, Advertising Research Foundation (ARF)
Using Simulated Data to Validate Marketing Mix Models – Jessica Nguyen, Meta
Mega or Micro? Influencer Selection Using Follower Elasticity – Ryan Dew, University of Pennsylvania
AI at the Front End of Innovation – Jessica Yankell, PepsiCo
InnoVAE: Generative AI for Patents, Innovations, and Firms – Dokyun Lee, Boston University
Recordings:
Livestream Shopping and Dynamic Customer Interactions – Xiao Liu, New York University
Applying “Explainable” AI: Using Theory to Understand AI Emotion Models – Hortense Fong, Columbia University
“Responsible AI” and Data Ethics – Miriam Vogel, EqualAI & JoAnn Stonier, Mastercard; Moderator Mary Purk, AI and Analytics for Business, The Wharton School
AI and the Future of Online Retailing Shopping – Nitzan Mekel-Bobrov, eBay
Using Simulated Data to Validate Marketing Mix Models – Jessica Nguyen, Meta
Mega or Micro? Influencer Selection Using Follower Elasticity – Ryan Dew, University of Pennsylvania
AI at the Front End of Innovation – Jessica Yankell, PepsiCo
InnoVAE: Generative AI for Patents, Innovations, and Firms – Dokyun Lee, Boston University
Session Summaries:
MSI 2023 Analytics Conference: Session Summaries