XIAO Ph.D. Candidate Office: GSIA 385A Tepper School of Business Cellphone: (412)251-9332 Carnegie Mellon University E-mail: [email protected] 5000 Forbes Avenue, Pittsburgh, PA, 15213 Website: http://www.andrew.cmu.edu/user/xiaoliu/

EDUCATION Ph.D. 2015 (Expected) Marketing, Carnegie Mellon University M.S. 2012 Industrial Administration (Marketing), Carnegie Mellon University B.S. 2010 Finance, Tsinghua University, China

RESEARCH INTERESTS High-tech Marketing, Consumer Financial Decision Making, Cloud Computing Dynamic Structural Models, Applied Game Theory

TEACHING INTERESTS Business Analytics, Pricing, Technology and Marketing Strategies

JOB MARKET PAPER “Overhaul Overdraft Fees: Creating Pricing and Product Design Strategies with Big Data” Dissertation Committee: Kannan Srinivasan (Chair), Alan Montgomery (Chair), Baohong , Param Vir Singh, Burton Hollifield

WORKING PAPERS  “A Structured Analysis of Unstructured Big Data Leveraging Cloud Computing,” 2014. (with Param Singh and Kannan Srinivasan) Major Revision at Marketing Science.  “An Empirical Analysis of Consumer Purchase Behavior of Base Products and Add-ons Given Compatibility Constraints,” 2014. (with Timothy Derdenger and Baohong Sun) Resubmitted to Marketing Science.  “The Impact of Grouping Product Assortment into Vices and Virtues on Consumer Self-Control,” 2013. (with Yuhuang and Joachim Vosgerau) Revise and Resubmit, at Journal of Consumer Research.

WORK IN PROGRESS  “New Product Introduction Signaling,” 2013. (with Kannan Srinivasan, Vineet Kumar and Timothy Derdenger)  “Mobile Payment Network Formation,” 2013. (with Kannan Srinivasan and Alan Montgomery)  “Omni-channel Competition,” 2013. (with Kannan Srinivasan and Alan Montgomery)

CONFERENCE PRESENTATIONS 2014 Marketing Science Conference, Atlanta, GA  Overhaul Overdraft Fees: Creating Pricing and Product Design Strategies with Big Data  A Structured Analysis of Unstructured Big Data Leveraging Cloud Computing 2012 Marketing Science Conference, Boston,  An Empirical Analysis of Consumer Purchase Behavior of Base Products and Add-ons Given Compatibility Constraints  Fraud Transactions under Seller Rating System: A Dynamic Analysis of Price and Quality Competition on Online Retailing Platform 2011 SCP 2011 Annual Winter Conference, Atlanta, GA  The Impact of Grouping Product Assortment Into Vices and Virtues on Consumer Self-Control

CONFERENCE PARTICIPATIONS 2014 AMA Sheth Doctoral Consortium, Evanston, IL 2014 Marketing Science Conference, Atlanta, GA 2013 Quantitative Marketing and Structural Econometrics Workshop, Durham, NC 2012 Quantitative Marketing and Economics Conference, Durham, NC 2012 Marketing Science Conference, Boston, MA 2011 Summer Institute in Competitive Strategy. Berkeley, CA 2011 Marketing Science Conference, Huston, TX 2011 Marketing and Industrial Organization Conference. New York, NY

TEACHING EXPERIENCE Instructor Marketing I (undergraduate), Summer 2013; Evaluation: 4.15/5.0 Teaching Assistant Marketing Management (MBA), Fall 2012, 2013 Technology Strategies (Undergraduate), Spring 2013 Auctions and Markets (Undergraduate), Fall 2012 Marketing Management (undergraduate), Summer 2012 Pricing Strategy (MBA), Spring 2011

GRANTS, HONORS AND AWARDS Winner of ISMS Doctoral Dissertation Proposal Competition 2014 Winner of MSI Alden G. Clayton Doctoral Dissertation Proposal Competition 2014 Dipankar and Sharmila Chakravarti Fellowship, 2014 AMA-Sheth Foundation Doctoral Consortium Fellow, Northwestern University, 2014

CMU GSA Conference Funding, 2014 INFORMS Marketing Science Doctoral Consortium Fellow, 2014 PNC Center for Financial Services Innovation Grant, 2013 Quantitative Marketing and Structural Econometrics Workshop Fellow, Durham 2013 William Larimer Mellon Fellowship, Carnegie Mellon University, 2010- 2015 INFORMS Marketing Science Doctoral Consortium Fellow, 2012 National Scholarship, Tsinghua University, 2009

GRADUATE COURSEWORK Course Instructor Marketing and Economics Analytical and Structural Models in Marketing Kannan Srinivasan Advanced Choice Models Baohong Sun Bayesian Statistics and Marketing Alan Montgomery Advanced Data Analysis Peter Boatwright Behavioral Foundations of Marketing Joachim Vosgerau Analytical Models of Marketing Kinshuk Jerath Dynamic Structural Models of Marketing and Economics Guofang Foundations of Consumer Behavior Cait Lamberton Behavioral Economics (audit) George Lowenstein Microeconomic Theory I Isa Hahalir Microeconomic Theory II Steve Spear Econometrics I: IV, MLE Fallaw Sowell Econometrics II: GMM Fallaw Sowell Econometrics III: Structural Models Robert Miller Econometrics IV: Small Sample GMM Fallaw Sowell Game Theory Applications Isa Hahalir Macroeconomics Christopher Sleet Economics of Contracts Lawrence Ales

Machine Learning, Computer Science and Statistics Cloud Computing Majd F. Sakr Machine Learning Eric Xing and Aarti Singh Applied Continuous Multivariate Analysis Vincent Q. Vu Intermediate Statistics Larry Wasserman Regression Analysis Kathryn Roeder Time Series Analysis Valarie Ventura Fundamentals of Programming Gregory Kesden Image, Video and Multimedia (audit)

State Space and Hidden Markov Models (audit) Jing

REFERENCES Kannan Srinivasan Rohet Tolani Distinguished Professor in International Business, H.J. Heinz II Professor of Management, Marketing and Information Systems Tepper School of Business, Carnegie Mellon University, Pittsburgh, PA 15213 Email: [email protected] Tel: (412)268-8840

Alan Montgomery Associate Professor of Marketing Tepper School of Business, Carnegie Mellon University, Pittsburgh, PA 15213 Email: [email protected] Tel: (412)268-2215

Baohong Sun Dean’s Distinguished Chair Professor of Marketing Cheung Graduate School of Business, 111 West 57th Street, Suite 418, New York, NY 10019 Email: [email protected] Tel: (646)627-7727

Abstracts of selected papers Overhaul Overdraft Fees: Creating Pricing and Product Design Strategies with Big Data (Job market paper) In 2012, consumers paid an enormous $32 billion overdraft fees. Consumer attrition and potential government regulations to shut down the overdraft service urge banks to come up with financial innovations to overhaul the overdraft fees. However, no empirical research has been done to explain consumers’ overdraft incentives and evaluate alternative pricing and product strategies. In this paper, we build a dynamic structural model with consumer monitoring cost and dissatisfaction. We find that on one hand, consumers heavily discount the future and overdraw because of impulsive spending. On the other hand, a high monitoring cost makes it hard for consumers to track their finances therefore they overdraw because of rational inattention. In addition, consumers are dissatisfied by the overly high overdraft fee and close their accounts. We apply the model to a big dataset of more than 500,000 accounts for a span of 450 days. To alleviate the computational burden of solving dynamic programming problems on a large scale, we combine parallel computing techniques with a Bayesian Markov Chain Monte Carlo algorithm. The Big Data equips us with a refined measure of consumer heterogeneity to compare new pricing structures and design targeted alerts. Our policy simulations show that alternative pricing strategies may increase the bank's revenue. Sending targeted and dynamic alerts to consumers can not only help consumers avoid overdraft fees but improve bank profits from higher interchange fees and less consumer attrition.

An Empirical Analysis of Consumer Purchase Behavior of Base Products and Add-ons Given Compatibility Constraints (with Timothy Derdenger and Baohong Sun) Despite the common practice of multiple standards in the high technology product industry, there is a lack of knowledge on how compatibility between base products and add-ons affects consumer purchase decisions at the brand and/or standard level. We recognize the existence of compatibility constraints and develop a dynamic model in which a consumer makes periodic purchase decisions on whether to adopt/replace a base and/or an add-on product. Dynamic and interactive inventory effects are included by allowing consumers to account for the -term financial implications when planning to switch to a base product that is incompatible with their inventory of add-ons. Applying the model to the consumer purchase history of digital cameras and memory cards from 2000 to 2004, we demonstrate that the inventory of add-ons significantly affects purchases of base products. This “lock-in” effect is enhanced when future prices of add-ons decrease. Interestingly, it is more costly for consumers to switch from Sony to other brands than vice versa. In four policy simulations, we explore the impact of alternative pricing and compatibility policies. For example, if Sony did not create its proprietary Memory Stick, the market share of its cameras would have been reduced by 6 percentage points.

A Structured Analysis of Unstructured Big Data Leveraging Cloud Computing (with Param Vir Singh and Kannan Srinivasan) Accurate forecasting of sales/consumption is particularly important for marketing as such information can be used to fine-tune marketing budget allocation and overall strategy. In recent years, online social platforms have produced unparalleled data on consumer behavior. However, there are two challenges that have limited the use of such data to get meaningful business insights in marketing. First, the data is typically in an unstructured format such as text, images, audio, video etc. Second, the sheer volume of data makes standard analysis procedures computationally unworkable. In this study, we combine methods from cloud computing, machine learning and text mining to illustrate how content from twitter can be effectively used for forecasting purposes. We conduct our analysis on a staggering volume of nearly two billion tweets. Our main findings highlight that, in contrast to just the surface level measures, such a volume of Tweets or sentiment in Tweets, the information content of the tweet and their timeliness improve forecasting accuracy significantly. Our method endogenously summarizes the information contained in Tweets by classifying it based on content matching across Tweets. The advantage of our method is that the classification of the Tweets is based on what is in the Tweets rather than preconceived topics that may not be relevant. We also find that in contrast to Twitter, the online search data (captured through google trends) is a very weak predictor of TV show demand. This is because while users tweet about the TV show before, during and after the show, but the TV show typically lags the show.

The Impact of Grouping Product Assortment Into Vices and Virtues on Consumer Self-Control (With Yuhuang Zheng and Joachim Vosgerau) How does grouping an assortment into vices and virtues—without labeling the groups as such—affect choice? We develop and test two competing hypotheses that grouping will decrease (H1a) or increase (H1b) choice share of virtues. In four experiments, we show the latter to be the case. Simply grouping choice options into virtues and vices without labeling them as such facilitates self-control because it makes the self-control conflict more salient. We demonstrate this self-control facilitating effect for hypothetical choices between magazines (experiment 1) and foods (experiment 2), and for real choices of soft drinks in two field experiments (experiments 3A and 3B). Finally, we show that making choices for others versus oneself moderates—and salience of the self-control goal mediates—the self-control facilitating effect (experiment 4).