Informs Annual Meeting 2017

TD27

INFORMS Houston – 2017

TD27

3 - A Big Data Driven Framework for Optimizing the Marketing Mix ROI using Fuzzy Neural Network Ajay Kumar, Indian Institute of Technology Delhi, Ajay Kumar, Room No - Eb-03,, Girnar Hostel, Iit Delhi , Hauz Khas, New Delhi, 110016, India, ajaytomar.dce@gmail.com, Ravi Shankar This study aims to investigate the contributions of promotional marketing activities as predictors and develops a big data driven framework for optimising marketing mix ROI using fuzzy neural network.Companies want to sense demand signals and shape future sales using historical demand, price, promotion and others economic factors so that they can react immediately to customer’s orders.This study improves the quality of sales forecasts using historical demand and sales data combined with advertising effectiveness,promotions and marketing events data,and provides important implications for practitioners as they can better understand how promotional marketing can influence product sales. 4 - Game-theoretic Analysis of Zero-rating Plans: The Importance of Context Considering the contrasting implications that zero-rating plans in the Internet domain are argued to have on society, we study the role of context (extent of digital divide) on these implications. Contrary to the belief that Internet Service Providers will allow only the content providers (CPs) with higher revenue generating capability to zero-rate their data and hence act as a gate-keeper to the Internet access, we find that the ISP will have the incentive to choose the different options (none, one or both CPs) in different context. We provide further direction to policymakers by analysing and comparing consumer and total surplus in each of the cases. 5 - Incorporating Diversity in Recommender Systems using Optimization Approaches Ibrahim Muter, Dr., University of Bath, Bath, United Kingdom, i.muter@bath.ac.uk Recommender systems have become one of the main components of web technologies, which helps people to cope with the information overload. These systems are based on the analysis of past behavior of users to develop user models that can filter items according to users’ likes and interests. Two of the most important metrics used to analyze the performance of these systems are accuracy and diversity of the recommendation lists which emerges in various forms such as aggregate diversity and individual diversity. In this study, we integrate these objectives using multi-objective optimization approaches and solve them using scalable optimization approaches. 350B Getting Funded by NSF: Proposal Preparation and the Merit Review Process Invited: NSF Invited Session Chair: Irina Dolinskaya, National Science Foundation, Herndon, VA, 20170, United States, irina.s.dolinskaya@gmail.com 1 - Getting Funded by NSF: Proposal Preparation and the Merit Review Process Irina Dolinskaya, National Science Foundation, Herndon, VA, 20170, United States, irina.s.dolinskaya@gmail.com, Georgia-Ann Klute So, you think you have a great research idea, now how do you get funding from the National Science Foundation (NSF) to do the work? A well-scoped and written proposal is instrumental to successful submission. This session targets junior faculty and researchers who might be new to NSF and describes detailed guidelines and practical advice for proposal preparation. The presenters will go over NSF review process and Intellectual Merit and Broader Impacts criteria, as well as share most common mistakes made by the PIs when submitting a proposal. Question-and-answer session will follow the presentation. TD26 Neena Pandey, Doctoral Candidate, Indian Institute of Management, Bangalore, India, neena.alok@gmail.com, Manaswini Bhalla

350C Social Media Analytics II Invited: Social Media Analytics Invited Session Chair: Karthik Kannan, karthikbabu.nk@scheller.gatech.edu 1 - Bias in Online Reviews: Variety and Atypicality in Online Gaming Pranay Jinna, Emory University, Atlanta, GA, United States, pjinna@emory.edu In this paper, I study self-selection bias in online reviews which can occur because the propensity of certain segments of users to opt in to review a product is different compared to other segments. Using a unique data set of 1.4 million individuals and their playing history across more than 3900 games, I find that high variety seeking consumers are more likely to post reviews compared to consumers who are low variety seeking. Further, these high variety seeking consumers have a different appeal for atypical games compared to low variety seeking consumers. I also find that users who have more friends on the platform are more likely to opt in to review. 2 - An Investigation of Information Technology and Complementarities in Innovation Patricia Angle, Georgia Institute of Technology, Atlanta, GA, United States, Patricia.Angle@scheller.gatech.edu, Chris Forman This study is an investigation into complementarities between IT and a firm’s intellectual property (IP) protection strategy, in the context of product innovation. An IP protection strategy is the way a firm protects its innovations from being appropriated for outside interests. Whether IT is complementary with an open strategy, such as the use of patents, or a closed strategy, such as trade secrets, is unclear. This early-stage research represents a novel investigation into the relationship between IT investments and a firm’s innovative strategy. 3 - Big Data Location Analytics - Predicting Purchase Location with Mobile Location Data Karthik Babu Nattamai Kannan, Scheller College of Business, Georgia Tech, 800 West Peachtree NW, Atlanta, GA, 30308, United States, karthikbabu.nk@scheller.gatech.edu, Sridhar Narasimhan We work with a leading national mobile carrier in the USA to explore how its customers choose the location of brick-and-mortar store to purchase or upgrade wireless service. We explore how understanding people’s travel patterns is helpful in choosing the optimal number of stores in a given market based on maximum coverage model. We explore literature on firm expansion and franchising to develop a machine learning based heuristic solution procedure (work in progress). 4 - Public Reaction to Supply Chain Glitches: A Twitter Sentiment Analysis David Wuttke, EBS.University, Burgstr. 5, Oestrich-Winkel, 65375, Germany, david.wuttke@ebs.edu, Christoph G. Schmidt, H. Sebastian Heese Supply chain glitches have become an inherent business risk and bad news spreads quickly through social media and can negatively affect sales. Our study investigates whether and how supply chain glitches evoke public reaction on Twitter in terms of attention and sentiment. We apply the event study method to 126 glitches, analyzing sentiments using machine learning. We find that public attention triples in response to glitches, and sentiments turn to negative. To examine the drivers of public reaction, we use attribution theory, and find that sentiment shifts are stronger when glitches result from stable, internal, and specific causes; thus helping managers to allocate resources more efficiently.

385