Information Technology, Marketing and Operations Research Seminars
Sequential IT Investment: Can the Risk of IT
Implementation Failure be Your Friend?
An extensive literature has studied the benefits for a firm to be the first to invest in innovative technologies such as Information Technologies (ITs). However, investment in innovative technologies faces high levels of uncertainty. How would such uncertainty affect a pioneering firm’s incentive to invest? Do late adopters benefit from information about the pioneer’s investment? In this paper, we investigate these questions in a context where firms engage in sequential investment in an innovative IT. This paper differs from prior literature in two aspects: First, IT adoption is nonexclusive and available to all client firms. Second, IT implementation can fail. In this case, a late adopter may have an information advantage since he can make investment decisions contingent on knowledge about the early adopter’s IT investment and implementation outcome. We use a standard Hotelling model of duopoly competition to examine firms’ incentives to sequentially invest in IT given the risk of IT implementation failure.
Our results show that the probability of IT implementation failure impacts firms’ investment incentives and profit through three distinct effects: the first-mover advantage mitigation effect, competition mitigation effect, and uncertainty-driven cost-based differentiation effect, although these three effects may drive the firms’ investment and profit in opposite directions. The follower’s knowledge about the leader’s IT investment level before making his own IT investment decision gives the leader a first-mover advantage and the follower a disadvantage. In contrast, the follower’s knowledge about the leader’s IT implementation outcome can benefit both the leader and the follower. Finally, we find that a spaced-out IT investment schedule in which the follower makes his investment decision after the Leader’s IT investment level and implementation outcome are known leads to the highest industry-wide IT investment and social surplus. We conduct a field experiment with the B2B retailer, providing salespeople with their own model’s price recommendations in real-time through the retailer’s CRM system, and allowing them to adjust their original pricing accordingly. We find that despite the loss of non-codeable information available to the salesperson but not to the model, providing the model’s price to the salesperson increases profits for treated quotes by 10% relative to a control condition. Using a counterfactual analysis, we show that while in most of the cases the model’s pricing leads to higher profitability, the salesperson generates higher profits when pricing for quotes or clients with unique or complex characteristics.
Mingdi Xin is an Assistant Professor of Information Systems at UCI Paul Merage School of Business.
Automating the B2B Salesperson Pricing Decisions: Can
Machines Replace Humans and When?
In a world advancing towards automation, we ask whether salespeople making pricing decisions in a high human interaction environment such as business-to-business (B2B) retail, can be automated, and when it would be most beneficial. Using sales transactions data from a B2B aluminum retailer, we create an automated version of each salesperson, that learns and automatically reapplies the salesperson’s pricing policy.
We conduct a field experiment with the B2B retailer, providing salespeople with their own model’s price recommendations in real-time through the retailer’s CRM system, and allowing them to adjust their original pricing accordingly. We find that despite the loss of non-codeable information available to the salesperson but not to the model, providing the model’s price to the salesperson increases profits for treated quotes by 10% relatively to a control condition. Using a counterfactual analysis, we show that while in most of the cases the model’s pricing leads to higher profitability, the salesperson generates higher profits when pricing for quotes or clients with unique or complex characteristics.
Accordingly, we propose a machine learning Random Forest hybrid pricing strategy, that automatically combines the model and the human expert and generates profits significantly higher than either the model or the salespeople.
Oded Netzer is a Professor of Business at Columbia Business School.
Analytics for Fine Wine: Realistic Prices
Earlier publications indicate that the price of young wines cannot be estimated accurately through weather conditions. Considering that a majority of wine trade occurs when the wine is young and even before it is bottled, determining realistic prices for these fine wines accurately is one of the most critical decisions. Surveys of leading wine merchants and industry experts exhibit significant departures from the realized market prices.
We develop a pricing model for fine wines using weather, market and expert reviews. The financial exchange for fine wines called Liv-ex has already adopted our estimated prices and tagged them as “realistic prices.” Our approach combines temperature, rainfall, market fluctuations and tasting expert scores and leads to accurate estimations that the wine industry has not seen before. Our study shows that higher temperatures, lower levels of precipitation, appreciation in the Liv-ex 100 index as a market indicator, and higher barrel scores increase market prices. We conduct a comprehensive set of robustness checks and show that the mean absolute deviation of actual market prices from our estimated prices is substantially smaller than any academic benchmark. Our realistic prices help create transparency in this highly opaque market. When compared with the realized prices, our realistic prices guide buyers (e.g., distributors, restaurateurs, merchants) in their purchase decisions as they can determine whether a wine is underpriced or overpriced.
Burak Kazaz is the Steven Becker Professor of Supply Chain Management at Syracuse University’s Whitman School of Management.
Empirical Research on Supplier Selection Criteria
Previous theoretical research assumes that offering price, product quality, delivery speed, and auxiliary services are the key factors that determine the attractiveness of a supplier to retailers. However, little knowledge is gathered via empirical studies to explain how supplier selection happens in reality. Our research wants to recover the supplier selection criteria by directly drawing empirical evidence from 20,000+ transactions between 440 distinct suppliers and thousands of retailers on a B2B platform with detailed transactional information and suppliers’ attributes.
Using a feature selection method named Gradient Boosted Decision Tree, we find in general, price and speed are more important than quality and service. However, the order changes with the volume of the transaction, the life cycle length of the product, and the expensiveness of the products. Moreover, we find only a handful of attributes matter in the decision. Consistent with information overloading theory, more information is not always helpful for carrying out more sophisticated choices. In the second phase of our study, the B2B platform updates its information display with price shown in a prominent position and fewer attributes displayed. We investigate how selection criteria under the price primacy information display are distinct from the previous criteria under the product primacy information display.
Keija Hu is an Assistant Professor of Operations Management at Vanderbilt’s Owen Graduate School of Management.
Which Brands Are Best Suited to Social Media Advertising? A Field Study of Social Media Advertising Effects on Consumer Attitudes
This research focuses on identifying which brand-related characteristics affect social media advertising effectiveness. We use a novel dataset from a large-scale global field experiment covering 110 brands running 235 Facebook and Instagram advertising campaigns between October 2015 and May 2017, and augment it with brand-generated posts on Facebook.
Using natural language processing, we analyze these posts to infer how “human-like” are the actions of different brands in the newsfeed environment. We find that social media advertising campaigns run by brands that behave in a more “human” manner are more likely to have a significant impact on upper-funnel mindset metrics, such as brand saliency.
Yakov Bart is an Associate Professor of Marketing at Northeastern’s D’Amore-McKim School of Business.
Channels of Impact: User reviews when quality is dynamic and
We examine the effect of managerial response on consumer voice in a dynamic quality environment. We argue that, in this environment, the
consumer is motivated to write reviews not by the possibility that the re-
views will impact the decisions of other consumers, but that the reviews
will impact the actions of the management and the quality of the service.
Dina Mayzlin is an Associate Professor of Marketing at USC’s Marshall School of Business.
Bayesian Analysis of Variance for Consumer Research
In this presentation I review the limitations of classical hypothesis testing with ANOVA and explain the advantages of taking a Bayesian approach. I introduce R software for hierarchical Bayesian ANOVA for the analysis of experiments. BANOVA alleviates the most important limitations of classical ANOVA, by including unobserved heterogeneity and allowing for dependent variables with a variety of distributions. The software accommodates (hierarchical) mediation, moderation, moderated mediation and floodlight analyses.
Michel Wedel is the Distinguished University Professor and PepsiCo Chair in Consumer Science at the University of Maryland’s Robert H. Smith School of Business.
Pricing Power of New Products: A Study of New Prescription
Pricing is one of the most critical strategic decisions that play a vital role in determining the success of new products. In this study, authors seek to study the link between the launch prices and pricing power of the product – the product’s ability to command an increased price without losing customers to competitors.
Rajdeep Grewal is the Townsend Family Distinguished Professor of Marketing at UNC’s Kenan-Flagler Business School and the editor-in-chief of the Journal of Marketing Research
A Methodology for Studying How Individuals Choose Locations in
Public Consumption Environments
Consumers often face public consumption environments (e.g., concerts, movie theaters, airplanes) in which they can tradeoff locational preferences with their desire to maintain suffcient personal space. After we define attributes for locational choices and the need for personal spaces, we introduce a Bayesian methodology that allows for the identification of the heterogeneous drivers of locational choices faced by consumers in such environments. We demonstrate the usefulness of the methodology with analyses of two scenarios illustrating how consumers choose seats at movie theaters and concert halls.
Simon Blanchard is an Assistant Professor of Marketing at Georgetown University’s McDonough School of Business.
Market Structure with the Entry of Peer-to-Peer Platforms: The
Case of Hotels and Airbnb
We study the entry of Airbnb in the accommodations industry to understand the determinants of flexible supply and its effects on travelers and incumbents.
Chiara Farronato is an Assistant Professor of Business Administration at Harvard Business School
FleetPower: Creating Virtual Power Plants in Sustainable Smart
Electric vehicles have the potential to be used as virtual power plants to provide reliable back-up power. This generates additional profits for carsharing rental firms, who rent vehicles by the minute. We show this by developing a discrete event simulation platform based on real-time locational information (GPS) of 1,100 electric cars from Daimlers carsharing service Car2Go in San Diego, Amsterdam, and Stuttgart.
Wolfgang Ketter is a Professor of Next Generation Information Systems at Erasmus University’s Rotterdam School of Management.
- 1 of 2