About Me:

I am an Assistant Professor of Marketing at the Wharton School of the University of Pennsylvania, where I am affiliated with Analytics at Wharton and the AI for Business initiative. My research explores the intersection of marketing, data science, and machine learning. At Wharton, I teach the class Data and Analysis for Marketing Decisions, for which I've been awarded the Wharton Teaching Excellence award.

Interested in my research or in collaborating? Feel free to email me at ryandew@wharton.upenn.edu. You can also find me on LinkedIn and on Google Scholar.

For more about my background, check out my CV.

Interested in probablistic machine learning, Bayesian computation, and marketing? I'm also the co-organizer of an interschool virtual reading group on these topics for junior faculty and Ph.D. students, which you can read more about here.

Research Overview:

I am an applied methodologist, interested in the intersection of modern probabilistic machine learning and marketing. I focus on problems in customer analytics, preference measurement, and design, with an eye to developing and applying flexible, interpretable, computational tools to drive insights in these domains. I also study how rich, unstructured data like text and images can be used in classical marketing contexts (like preference measurement), and in new contexts (like visual branding and product design) which previously have been difficult to study from a data-driven perspective. Methodologically, I am interested in Bayesian nonparametrics, Bayesian computation, deep generative models, and representation learning. I'm honored to have received several awards for my research, including the 2018 INFORMS Society for Marketing Science Doctoral Dissertation Award, and the 2018 Marketing Section of the American Statistical Association's Doctoral Research Award, as well as to have been a finalist for the 2019 Frank M. Bass award.


Publications:

• Bayesian Nonparametric Customer Base Analysis with Model-based Visualizations
  Ryan Dew and Asim Ansari
  Marketing Science, 2018
  [Show Abstract] [Paper] [Code Notebook] [Replication Data] [Stan Code]
  

  Modern marketers are responsible for understanding and managing customer spending behavior across many different products. Dynamics in spending result from both predictable customer-level effects, which are characterized by interpurchase time, customer lifetime, and past purchase frequency, as well as calendar time effects, which are driven by managerial actions such as product changes and promotions, and by general trends and random shocks. Understanding these dynamics in spending is further complicated by a lack of knowledge of all of the factors that influence spending for a given product, a problem exacerbated in large multiproduct firms by information asymmetries that can exist between the product teams that execute marketing actions and the marketing analytics team responsible for customer base analysis. A comprehensive understanding of customer base dynamics therefore requires a modeling framework that flexibly integrates both known and unknown calendar time determinants of spending with the individual-level effects that robustly predict spend activity. In this paper, we develop a Bayesian nonparametric framework based on Gaussian process priors to understand and predict customer spending. Our model separates out calendar time effects from individual-level dynamics by modeling both sets of factors as unknown latent functions that jointly determine spend propensity. The primary output of our Gaussian Process Propensity Model (GPPM) is a set of estimated curves that provides a visual and easily comprehensible representation of purchasing dynamics, which we call the model-based dashboard. We illustrate the utility of our modeling framework on data from two popular free-to-play mobile video games. We show how the GPPM's model-based dashboard can be useful for assessing patterns and disruptions in spending. We also show how the GPPM exhibits superior forecasting ability compared to existing customer base analysis benchmarks, including hazard and buy-till-you-die models.



• Modeling Dynamic Heterogeneity using Gaussian Processes
  Ryan Dew, Asim Ansari, Yang Li
  Journal of Marketing Research, 2020
  [Show Abstract] [Paper]
  

  Marketing research relies on individual-level estimates to understand the rich heterogeneity that exists in consumers, firms, and products. While much of the literature focuses on capturing static cross-sectional heterogeneity, little research has been done on modeling dynamic heterogeneity, or the heterogeneous evolution of individual-level model parameters. In this work, we propose a novel framework for capturing the dynamics of heterogeneity, using individual-level, latent, Bayesian nonparametric Gaussian processes. Similar to standard heterogeneity specifications, our Gaussian Process Dynamic Heterogeneity (GPDH) specification models individual-level parameters as flexible variations around population-level trends, allowing for sharing of statistical information both across individuals and within individuals over time. This hierarchical structure provides precise individual-level insights regarding parameter dynamics. We show that GPDH nests existing heterogeneity specifications, and that not flexibly capturing individual-level dynamics may result in biased parameter estimates. Substantively, we apply GPDH to two problems: understanding preference dynamics, and modeling the evolution of online reviews. Across both applications, we find robust evidence of dynamic heterogeneity, and illustrate GPDH's rich managerial insights, with implications for targeting, pricing, and market structure analysis.

Working Papers:

• Letting Logos Speak: Leveraging Multiview Representation Learning for Data-Driven Logo Design
  Ryan Dew, Asim Ansari, Olivier Toubia
  Revision invited at Marketing Science.
  [Show Abstract] [Working Paper] [Explore Our Data] [Personality-based Logo Generator]
  

  Logos serve a fundamental role in branding as the visual figurehead of the brand. Yet, due to the difficulty of using unstructured image data, prior research on logo design has been largely limited to non-quantitative studies. In this work, we explore logo design from a data-driven perspective. In particular, we aim to answer several key questions: first, to what degree can logos represent a brand's personality? Second, what are the key visual elements in logos that elicit brand and firm relevant associations, such as brand personality traits? Finally, given text describing a firm's brand or function, can we suggest features of a logo that elicit the firm's desired image? To answer these questions, we develop a novel logo feature extraction algorithm, that uses modern image processing tools to decompose unstructured pixel-level image data into meaningful visual features. We then analyze the links between firm identity and the features of logos through a deep, multiview generative model, which links visual features of logos with textual descriptions of firms and consumer ratings of brand personality by learning representations of brand identity. We apply our modeling framework on a dataset of hundreds of logos, textual descriptions from firms' websites, third party descriptions of firms, and consumer evaluations of brand personality to explore these questions.



• Detecting Routines in Ride-sharing: Implications For Customer Management
  Ryan Dew, Eva Ascarza, Oded Netzer, and Nachum Sicherman
[Show Abstract] [Working Paper]


Routines are central to consumer behavior in many industries, including ride-sharing, where consumers may use the same app to take the same trips on a regular basis. While prior work has established the importance of repeat behavior for marketing, little work has been done to understand the implications of routines, which we define as repeated behavior with a distinct, recurring, temporal structure. Partly, this lack of research stems from the statistical problem of estimating routines. In this paper, we propose a new approach to measuring routine usage, which we apply in the context of ride-sharing. Specifically, we model usage of the platform as an individual-level inhomogeneous Poisson point process, where the rate of usage is determined partly by a Bayesian nonparametric Gaussian process. In estimat- ing this rate function, we leverage a unique cyclical kernel structure, that allows for precise estimation of recurrent behavior. We then use this model to estimate individual-level rou- tines in usage of a ride-sharing service. We show that more routine users tend to be more valuable customers, with high individual-level “routineness” being significantly associated with higher future usage and lower churn rates.



• Decomposing The Impact of a Free Cancellation Program on Customer Booking Behavior
  Yuhao Fan, Ryan Dew, Eric T. Bradlow, Peter Fader
[Show Abstract] [Request Paper]


Despite the recent surge in popularity of free cancellation programs across many business settings, the impact of a free cancellation program on customers’ behavior and firm profits remains unclear. Previous research has found that the implementation of a free cancellation program encourages customers to repeat purchase more frequently, spend more on each trip, but also cancel their reservations more often. However, we do not know the relative magnitude of each effect, or which may dominate in terms of the overall impact on firm profit. We investigate this question empirically using data from a hostel booking platform, and a Bayesian semiparametric model that lets us nonparametrically decompose the impact of the free cancellation on customers’ repeat booking propensity, booking timing, spend and cancellation amounts, and propensity to churn. Our model combines both Bayesian nonparametric Gaussian processes to assess intertemporal effects of the program, and Dirichlet processes to capture the rich heterogeneity that exists in this setting. We find that the effect of the program on existing customers varies tremendously across different customer segments, with muted short-run effects for loyal customers, but strong, positive CLV implications. Ultimately, we find the program was beneficial for the platform.

Research in Progress:

• Preference Measurement with Unstructured Data, with Applications to Adaptive Onboarding Surveys
  Ryan Dew
[Show Abstract]


A common problem in recommendation engines is the cold start problem: how can we make a recommendation to a new customer, without any prior purchase data? Such problems are particularly salient for increasingly common online subscription businesses, where initial recommendations can shape whether potential customers decide to subscribe, and how their preferences evolve subsequently. The need to assess a new customers' preferences quickly, and without prior purchase data, has led to the increasing prevalence of customer onboarding surveys, wherein companies ask potential or current customers a series of questions aimed at understanding their preferences, without having observed any purchasing. While such onboarding surveys are a relatively recent development in e-commerce, the idea of learning the most information about a customer’s preferences as possible using the fewest questions has been studied extensively in “offline” marketing research, in the context of adaptive conjoint analysis. In this work, I bridge these two domains using a combination of representation learning for unstructured data, and Bayesian optimization for on-the-fly estimation of preferences. I apply this framework both in the context of an on-boarding survey for an online subscription business, and in the context of traditional preference measurement.



• A General, Kernel-based Framework for Capturing Cross-Category Choice Dynamics
  Ryan Dew and Yuhao Fan
[Show Abstract]


Understanding individual customers' sensitivities to prices, promotions, brand, and other aspects of the marketing mix is fundamental to a wide swath of marketing problems, including targeting and pricing. Companies that operate across many product categories have a unique opportunity, insofar as they can use purchasing data from one category to augment their insights in another. Such cross-category insights are especially crucial in situations where purchasing data may be rich in one category, and scarce in another. An important aspect of how consumers behave across categories is dynamics: preferences are not stable over time, and changes in individual-level preference parameters in one category may be indicative of changes in other categories, especially if those changes are driven by external factors. Yet, despite the rich history of modeling cross-category preferences, the marketing literature lacks a framework that flexibly accounts for both the cross-category aspect of preferences, and individual-level dynamics. In this work, we propose such a framework, leveraging individual-level, latent, multi-output Gaussian processes to build a nonparametric Bayesian choice model that allows information sharing about preference parameters across customers, time, and categories. We apply our model to a data set of grocery purchases, and show that our model detects interesting dynamics of customers' price sensitivties across multiple catergories, and that the ability to detect preference dynamics that persist across categories is important in being able to make predictions across categories, especially in situations when data is scarce.



• An Investigation into Video, Hashtag, and Challenge Success on TikTok
  Zijun Tian, Ryan Dew, and Raghu Iyengar
[Show Abstract]


In this project, we explore the empirical regularities governing the diffusion of content on the recently popular social media platform TikTok, including video and hashtag-level drivers of content popularity. TikTok is a unique social media platform focused on the sharing of short videos. Similar to social media platforms like Twitter, TikTok features hashtags that organize user generated content. However, unlike other social media platforms, TikTok also features a special type of hashtag, the challenge, that encourages users to generate (rather than merely share) content matching a particular theme. Combining both classical ideas from the literature on diffusion on social networks with cutting-edge methods from machine learning and computer vision, we explore how these challenges evolve in popularity, and what features of videos dictate their success.