SPEAKER

Ms. Hortense Fong
Ph.D. Candidate in Quantitative Marketing
Yale School of Management
Yale University

ABSTRACT

Music is used extensively to evoke emotion throughout the customer journey. This paper develops a theory-based, interpretable deep learning convolutional neural network (CNN) classifier-MusicEmoCNN-to predict the dynamically varying emotional response to music. To develop a theory-based CNN, we first transform the raw music data into a-mel spectrogram-a format that accounts for human auditory response as the input into a CNN. Next, we design and construct novel CNN filters for higher order music features that are based on the physics of sound waves and associated with perceptual features of music, like consonance and dissonance, which are known to impact emotion. The key advantage of our theory-based filters is that we can connect how the predicted emotional response (valence and arousal) are related to human interpretable features of the music. Our model outperforms traditional machine learning models and performs comparably with state-of-the-art black box deep learning CNN models. Our approach of incorporating theory into the design of convolution filters can be taken to settings beyond music. Finally, we use our model in an application involving digital advertising. Motivated by YouTube’s mid-roll advertising, we use the model’s predictions to identify optimal emotion based ad insertion positions in videos. We exogenously place ads at different times within content videos and find that ads placed in emotionally similar contexts are more memorable in terms of higher brand recall rates.

SPEAKER

Ms. Fanglin Chen
Ph.D. Candidate in Marketing
NYU Stern School of Business
New York University

ABSTRACT

Studying competition and market structure at the product level instead of brand level can provide firms with insights on cannibalization and product line optimization. We introduce Product2Vec, a method based on representation learning, to study product-level competition when the number of products is large. The proposed model takes shopping baskets as inputs and, for every product, generates a low-dimensional vector that preserves important product information. Using these product vectors, we present several findings. First, we show that these vectors can recover analogies between product pairs. Second, we create two measures, complementarity and exchangeability, that allow us to determine whether product pairs are complements or substitutes. Third, we combine these vectors with traditional choice models to study product-level competition. To accurately estimate price elasticities, we modify the representation learning algorithm to remove the influence of price from the product vectors. We show that, compared with state-of-the-art choice models, our approach is faster and can produce more accurate demand forecasts and price elasticities. Fourth, we present two applications of Product2Vec to marketing problems: 1) analyzing intra- and inter-brand competition and 2) analyzing market structure. Overall, our results demonstrate that machine learning algorithms, such as representation learning, can be useful tools to augment and improve traditional marketing methods.

SPEAKER

Dr. Hui Li
Associate Professor of Marketing
Tepper School of Business
Carnegie Mellon University

ABSTRACT

This paper examines the impact of Airbnb on the local rental housing market. Airbnb provides landlords an alternative opportunity to rent to short-term tourists, potentially leading some landlords to switch from long-term rentals, thereby affecting rental housing supply and affordability. Despite recent government regulations to address this concern, it remains unclear how many and what types of properties are switching. Combining Airbnb and American Housing Survey data, we estimate a structural model of property owners’ decisions and conduct counterfactual analyses to evaluate various regulations. We find that Airbnb mildly cannibalizes the long-term rental supply. Cities where Airbnb is more popular experience a larger rental supply reduction, but they do not necessarily have a larger percentage of switchers. Affordable units are the major sources of both the negative and positive impacts of Airbnb: they cause a larger rental supply reduction, which harms local renters; they also create a larger market expansion effect, which benefits local hosts who own affordable units and may be less economically advantaged. Policy makers need to strike a balance between local renters’ affordable housing concerns and local hosts’ income source needs. We also find that imposing a linear tax is more desirable than limiting the number of days a property can be listed. We propose a new convex tax that imposes a higher tax on expensive units and show that it can outperform existing policies in terms of reducing cannibalization and alleviating social inequality. Finally, Airbnb and rent control can exacerbate each other’s negative impacts.

Marketing Seminar

Speaker:

Mr. Piyush Anan
Ph.D. Candidate
S C Johnson Graduate School of Management
Cornell University

Abstract:
E-cigarette usage has grown significantly in the last few years, with special health concerns for underage usage. In response to this growing public health crisis, various states have enacted higher taxes to deter usage. Since underage usage is illegal, it is difficult to find data on their usage, and hence estimate the impact of these taxes. We exploit within – and across-state variation of publicly available 388,593 user-posted images on social media from Jan 2016 – Dec 2018 to measure the impact of greater taxes on underage posting behavior. This posting behavior is a rough proxy for influencing and normalizing behavior, and possibly for consumption behavior among underage population. We use an ensemble of image analysis methods – Mask R-CNN (He et al. 2017) and Aggregated Residual Neural Networks (Xie et al. 2017) to detect underage posters, and other demographics. Importantly, we develop methods to measure disguised posting of usage images, given their purported utilization by underage users. By using generalized synthetic controls (Xu 2017), we find that the states with higher taxes – Pennsylvania and California-saw a decline in underage e-cigarette posts. California’s decline is preceded by an increase in disguised posting, and Pennsylvania’s decline is accompanied by increased engagement with the underage posts. States with lower taxes – Kansas and West Virginia- saw no changes in underage posting. We also examine tax impact on posting by gender and race, given concerns of unequal health outcomes for particular groups. Our proposed approach of measuring e-cigarette posting behavior using social media images is likely to be of interest to regulators, managers, and businesses.

Speaker:

Dr. Kitty Wang
Assistant Professor
Bauer College of Business
University of Houston

Abstract:

We use data from a large U.S.-based specialty retail brand and a similar control brand before and after an involuntary revelation of the focal brand’s political position to study if and how corporate political advocacy (CPA) affects sales. We find that, on average, total sales of the focal brand do not change significantly after the event relative to the control brand. However, sales increase in places where the local political preference aligns with the focal brand’s position and decrease in places where the local political preference misaligns with the focal brand’s position. The change in sales after the event ranges from –26.8% (–20.4%) to 69.7% (57.6%) with a mean of 3.73% (5.51%) for sales dollar amount (quantity) across locations. We also find that the change in the customer base rather than basket size drives the effect of CPA. In addition, changes in online sales drive the change in total sales, and there is no qualitative difference between the shift of purchase of conspicuous and inconspicuous products after the event. These data patterns suggest that consumers’ reactions to CPA are motivated by their intrinsic need to support political ideologies, rather than the need to signal their political ideologies to others.

Speaker:

Dr. Peter Zubcsek

Assistant Professor of Economics

Coller School of Management

Tel Aviv University

ABSTRACT

Scan and Go (SAG) is a form of retail self-checkout. Shoppers using SAG scan items as they shop and pay at checkout without having to rescan the cart contents. Many retailers are adopting SAG systems to simultaneously increase customer satisfaction and lower store labor costs. However, criminology research has linked SAG to increased shopper theft, and has suggested to offset inventory losses in ways that each increase store staff, reducing the savings from SAG. We propose to extend the scope of tools retailers can use to reduce SAG shoplifting. Based on the equity theory and price fairness literatures, we posit that shoppers who purchase more discounted items will perceive the store as having treated them more fairly, and will therefore steal less. We confirm our prediction on data from a supermarket chain that uses SAG, finding that discounts may generate indirect savings (by reducing theft) corresponding to a significant proportion of retailers’ operating margins. Our findings carry important implications for grocery retailers, whose profitability has been increasingly under pressure from various emerging forms of competition.

Marketing Seminar

Speaker:

Mr. Wei MIAO
Ph.D. Candidate in Quantitative Marketing
NUS Business School
National University of Singapore

Abstract:

The ride-hailing industry is a critical pillar of modern transportation infrastructure and generates massive amounts of revenue each year. Due to spatial mismatch and search friction, the conventional taxi business model, which is based on street hailing, leads to substantial matching inefficiency. With the advent of geolocation-based mobile apps, ride-hailing firms can now effectively bridge demand and supply via digitalized matching technology and benefit from more flexibility in setting their pricing menus. In this paper, we analyze an exogenous event that the largest taxi operator in Singapore added an origin-destination-based flat fare option to its existing metered fare option. We empirically examine the effect of flat fare pricing vis-à-vis metered pricing on the outcome of this two-sided marketplace. Specifically, we model taxi drivers’ location choices as a dynamic spatial oligopoly game in which vacant drivers decide where to search for passengers, given the search behaviors of their competitors, in the presence of trip uncertainties. We leverage the large number of agents in the taxi industry and solve for the Oblivious Equilibrium (Weintraub, Benkard, and Van Roy 2008), in which each taxi driver’s policy function is based on their beliefs about the transition of average industry states. We then plug supply estimates into the demand system and recover demand parameters with a parametric aggregate-level matching function that accounts for matching inefficiency on street hail trips. We find that drivers are risk-averse on flat fare trips, especially during peak hours when trip uncertainty is higher, and riders’ risk aversion on metered trips also confers a risk premium on the flat fare pricing option. Finally, we run two counterfactual experiments to quantify the economic value of risk aversion for both riders and drivers, and evaluate the benefit of a booking system that enables flat fares. Our findings have important managerial implications for the rapidly expanding ride-hailing industry.

Marketing Seminar

Speaker:

Miss Zhenqi LIU
Ph.D. Candidate in Economics
Department of Economics
University of Pennsylvania

Abstract:

This paper studies the role of market structure in regulatory enforcement through a unique empirical example: censorship via content removal by three live-streaming platforms in China. Adopting an event study approach, this paper shows that platforms of different sizes censor a different number of keywords with notably different delays and their traffic declines after censorship. This paper then develops a model where the platform’s profit depends on its own censorship action as well as that of its competitors, induced by the switching behavior of users with heterogeneous preferences for censorship. By complying with the government’s censorship request, platforms may lose users who prefer to evade censorship by switching out. By not censoring, platforms incur a cost imposed by the government that is positively correlated with their sizes, but it also allows them to attract new users from their competitors that do censor. The model predicts that when the political pressure is sufficiently high and platforms are of similar size, they are less likely to censor as the number of competitors increases. If platforms are highly asymmetric in size, small platforms have strong incentives to differentiate themselves from their big competitors by not censoring, while big platforms find it costly not to censor. When the market becomes more concentrated, the behavior of big platforms dominates and as a result, users experience more censorship.

Marketing Seminar

Speaker:

Miss Jisu CAO
Ph.D. Candidate in Economics
Department of Economics
University of Southern California

Abstract:

The objective of this paper is to study consumer purchase timing and product returns for daily deal e-commerce where products are often sold in a short window of time (usually one to three days). Leveraging a unique proprietary data set from a leading Chinese daily deal website, we find two interesting patterns: (1) consumers generally buy earlier rather than later in sales events; and (2) product return rates are higher for consumers who purchase earlier. These empirical patterns may suggest a potential problem to daily deal merchants: a sales promotion to encourage consumers to buy earlier may actually increase consumer return probabilities and possibly hurt profits. To understand such tradeoffs, we develop an integrative model of consumer purchase timing and product return decisions. In a post-purchase stage, consumer knowledge of the product fit gets realized, and the consumer can return the product with some cost. In a purchase (order) stage, the consumer makes the purchase decision based on her expected utility considering the return probability. A forward-looking consumer solves an optimal stopping problem for a finite-horizon time period game to decide when to purchase in a sales event. Delaying purchase allows the consumer to see newly posted offers and have more time to consider her purchase. We estimate our structural model using a panel data of the purchase and return histories of 5,000 consumers of women’s clothing from January to June 2017. We find that the proposed model fits the data well and that competition significantly increases a consumer’s probability of buying later. In the counterfactual analysis, we adjust the product price over days of a sales event and compare the merchants’ profits under different pricing schedules. Our counterfactual results reveal important managerial insights and can help daily deal merchants select a pricing schedule to improve profit.

From Free to Paid: Testing Monetization Strategies for a Free Non-advertising-based Service

Speaker:

Mr. Jingcun CAO
Ph.D. Candidate in Marketing
Kelley School of Business
Indiana University

ABSTRACT

The mobile app market has expanded rapidly in the last decade. However, mobile app firms of non-advertising-based services face a significant challenge when trying to monetize their free services, due to low purchase conversion rates and high churn rates. In practice, these developers usually adopt one of two monetization strategies: (1) a soft-landing strategy, with limited free service provided to current users when it starts to charge, or (2) a hard-landing strategy in which all free services are terminated and only paid users retain access to the service. Which of these strategies may provide the best economic outcome for the developers is not clear. Developers are also uncertain about whether to add exclusive secondary offerings (unrelated to the core benefits) to the app for paid users. Existing theoretical and empirical literatures, however, do not offer guidance on the above choices. In this study, we employ a series of large-scale randomized field experiments to test the causal effects of providing limited free services to users and offering exclusive secondary offerings to those who pay when they convert to paid subscriptions. We also examine the interaction effects of these offers. Results suggest app users are more willing to pay in the hard-landing condition than in the soft-landing case regardless of whether free content or free time is offered. Additionally, exclusive secondary offerings hurt the conversion rate. Further, a positive interaction effect exists between providing limited free services and offering exclusive secondary offerings. Our research provides important theoretical implications to marketing researchers, and actionable managerial implications to mobile-app managers on monetization strategies and customer relationship management for non-advertising-based apps.