Ph.D. in Information Systems
2013 - 2018 (expected)

H. John Heinz III College,
Carnegie Mellon University

Post Graduate Diploma in Management
(equivalent to M.B.A.)
2008 - 2010

Indian Institute of Management Kozhikode, India

Bachelor of Technology, Electrical Engineering

Indian Institute of Technology Madras, India


Research Interests

  • Topics: Online Advertising, Privacy, Mobile Economy, Digital Word-of-Mouth, User-generated Content, Platforms & Marketplaces, Pro-social Giving
  • Methodologies: Causal Inference Methods including Field, Natural & Quasi-Experiments; Machine Learning; Game Theory

Working Papers

  1. Irrelevant Advertising: How More Consumer Tracking can Make Ads Less Effective, with Rahul Telang
  2. The online activity of consumers is routinely tracked, not just by visited websites, but also by an increasing number of third-parties who embed their tracking cookies within various websites. This tracked information is sold to advertisers who then use it to identify and target potential new consumers with their ads. In this paper, we build a game theoretic model to show that the presence of such third-party information can not only reduce advertiser profits, but also create a dead-weight loss under certain conditions.

Work in Progress

  1. How do Ratings impact Online Marketplaces? - Evidence from a Quasi-Experiment, with Hui Li & Rahul Telang
  2. Online ratings have been shown to have an influence on consumer choices. However, it is not well understood whether the design of the rating system has an impact on consumer behavior. To answer this question, we designed and conducted a quasi-experiment in partnership with a large online marketplace, where we modified the method of calculating the displayed rating for each listing. Using a differences-in-differences estimation approach, we find that total sales on the platform, as well as number of new customers joining the platform both increase significantly. We propose two distinct mechanisms to explain these effects - the rating effect and the rank effect, that work on different ends of the popularity spectrum. We also find that the effects on sales are mostly driven by new customers.

  3. Privacy-Aware Prediction using Modified Lasso, with Rahul Telang
  4. Most consumers report feeling that they have little information about how their information is tracked and utilized. In this paper, we propose a new methodology that balances the trade-off between how informative a variable is with how privacy-intrusive it is perceived to be by consumers. We do this by modifying the Lasso, a model selection method. Instead of applying equal penalty weights for each variable included in the model, we instead weight the penalties in proportion to how privacy-intrusive the variable is perceived to be by consumers.

  5. Effect of Service Quality on Product Ratings
  6. In this paper, we investigate whether the service quality of a platform spills over into the consumer's rating of product quality. We partner with a large online marketplace which collects distinct consumer ratings on service and product quality after every transaction. We exploit an anomaly in the marketplace algorithm which exogeneously and systematically estimates longer delivery times for some consumers, implying poorer service quality. Actual delivery times are much shorter than predicted times, and consumers experience that as a positive service quality shock. We find that this service quality shock not only impacts service quality ratings, but also increases product ratings.

  7. Impact of Private-Label Entry on Competing Retailers
  8. Online platforms, such as Amazon, have recently expanded their own catalogue of private-labels into categories such clothing, electronics, household essentials and more. The extent to which such private-labels cannibalize sales from existing brands is an open question. On the other hand, the availability of such private-labels can attract new customers to the platform, which might have a positive spill-over effect on existing sellers on the platform. We investigate the impact of such entry of private-labels by using entry of a private-label from a hyperlocal online marketplace. We exploit the hyperlocal nature of the marketplace, and use consumers who cannot access the private-label as the control group. We find the causal effect of private-label entry through a differences-in-differences estimation strategy.

  9. Capturing Empathy: Using Deep Learning and Experiments to Stimulate Social Giving
  10. Online platforms have enabled peer-to-peer charitable giving by matching potential donors with social causes. On a typical charitable giving platform, each individual recipient or cause has a dedicated webpage with textual description and pictures. In this project, we attempt to first use Deep Learning methods to predict what features of the textual description and pictures evoke charitable giving by donors. Next, we run a field experiment in which the descriptions of a selection of new causes are modified based on the Deep Learning extracted features. We measure and report their causal effect on social giving.

Conference Presentations

  1. "What is a Cookie Worth? Ad Effectiveness versus Consumer Privacy", Workshop on Information Systems and Economics (WISE) , Fort Worth, Texas, 2015,
    Winner - Best Student Paper Award
  2. "What is a Cookie Worth?" NBER Summer Institute, Economics of Digitization , Boston, 2015
  3. "What is a Cookie Worth? Ad Effectiveness versus Consumer Privacy", Conference on Information Systems and Technology (CIST) , Philadelphia, 2015
  4. "What is a Cookie Worth? Ad Effectiveness versus Consumer Privacy", The Economics of Information and Communication Technologies Conference , Paris, 2015
  5. "What is a Cookie Worth?", Sixth Annual Conference on Internet Search and Innovation, Searle Center , Northwestern University, 2015
  6. "Consumer Tracking and Targeted Advertising", Workshop on Information Systems and Economics (WISE) , Dublin, Ireland, 2016
  7. "Consumer Tracking and Competition in Targeted Advertising", American Marketing Association Summer Conference , San Francisco, 2017
  8. "Consumer Tracking and Competition in Targeted Advertising", Conference on Information Systems and Technology (CIST) , Houston, 2017
  9. "How do Ratings impact Online Marketplaces? - Evidence from a Quasi-Experiment", Conference on Digital Experimentation (CODE) , Boston, 2017


I'm interested in teaching Business Analytics, Data Science and Data Mining courses for undergraduate and graduate students.
I've been a Teaching Assistant for graduate students at Heinz College, CMU for the following courses:

  • Business Analytics for Managers
  • Business Intelligence & Data Mining
  • Managing Disruptive Innovation
  • Marketing Management
  • Economic Analysis
  • Accounting & Finance Fundamentals