Research Seminars

Seeking and Avoiding Choice Closure

Marketing

Speaker: Yangjie GU
Assistant Professor , Tilburg University, The Netherlands

11 December 2014 - Building T, Room 201 (T201) - From 10:30 am to 12:00 pm


Achieving a sense of finality with past decisions, or reaching closure with one’s choices, is generally regarded as beneficial to subjective wellbeing. We instead investigate circumstances under which reaching closure about a choice prevents satisfaction maximization: when post-choice information characterizes a decision outcome as inferior, choice closure inhibits unfavorable comparisons and enhances satisfaction; but when it characterizes this outcome as superior, choice closure inhibits favorable comparisons and fails to enhance satisfaction. Building on research showing that closure can be externally facilitated, we demonstrate that visual cues can trigger a sense of finality. To maximize satisfaction, consumers should seek such triggers when faced with an inferior outcome and avoid them when faced with a superior outcome. However, we show that participants decide to employ or avoid choice-closure triggers in ways that fail to enhance satisfaction, and that nudging them against their default inclinations may be needed in regulating post-choice satisfaction.

Intertemporal Discounting and Reference Durations

Marketing

Speaker: Florian STAHL
Professor of Marketing , University of Mannheim, Germany

9 December 2014 - Room T201 - From 2:00 pm to 3:30 pm


Intertemporal Discounting and Reference Durations

by Florian Stahl, Raghuram Iyengar, Yuxin Chen and Andreas Herrmann

Products and services are increasingly offered with contracts of different length. Consumers' choice of a specific contract involves an intertemporal decision, as they have to discount future utility. Given the long duration of some contracts, it is likely that consumers' instantaneous utility for a service is time dependent. We study individuals' discounting behavior allowing for changes in instantaneous utility. We gather experimental data from matching tasks and identify discounting patterns using a latent change-point model. Our results show that models with change-points fit individuals' discounting pattern significantly better than models without. There is also conceptual superiority of including change-points as these are highly correlated with durations that consumers may consider in their decisions (e.g., time till graduation). Interestingly, individuals' discounting pattern is consistent with exponential discounting in the absence of change-points but follows a hyperbolic discounting pattern when allowing consumers' instantaneous utility to change over time. Using a novel monetarily incentivized experimental setting, we show that individuals are not able to discount future outcomes without a bias towards the present even when they have an incentive to do so.

Unveiling Hidden Markov Models in Marketing, One-day Workshop Labex-Ecodec

Marketing

Speaker: Eva Ascarza and Oded Netzer
Assistant Professor of Marketing and Associate Professor of Business , Columbia University

19 September 2014 - Room T033 - From 8:30 am to 6:00 pm


Unveiling Hidden Markov Models in Marketing

Customers or the business environment often transition over time from one mode of behavior to another. For example, a customer may transition from a more positive relationship with the firm to a less positive one due to exposure to an attractive competitor. However, managers often do not observe the customer’s buying behavior state, neither the transitions from one state to another. Rather, they need to infer such changes from the customer’s observed behaviors (e.g., purchases, transactions, searches). Hidden Markov models (HMMs) are valuable tools for such situations. More generally, HMMs are useful for situations in which the marketer or researcher is interested in identifying a (dynamic) latent state of the world from a series of (possibly noisy) observations. Recently, these models have been widely applied to marketing problems. In marketing, firms are often interested in understanding their customers’ latent buying behavior state and assessing how the firm can use its marketing levers to move customers to a more favorable state to the firm.

In this workshop we will discuss what are hidden Markov models, what are their advantages and perils and how should one go about applying such models to marketing data. We will demonstrate these topics using example of applications of these models in marketing and related fields. The workshop will also include hands on estimation of hidden Markov models using R. No pre-existing knowledge of R will be assumed, though participants are encouraged to install the R statistical software (freely available from http://www.r-project.org/) on the laptops prior to the workshop.

How firms can go wrong by offering the right service contract: Evidence from a field experiment

Marketing

Speaker: Eva Ascarza
Assistant Professor of Marketing , Columbia Business School

18 September 2014 - Room T025 - From 1:00 pm to 2:30 pm


"How firms can go wrong by offering the right service contract: Evidence from a field experiment”


Eva Ascarza (with Raghuram Iyengar, and Martin Schleicher)


Abstract:

Service firms typically offer customers a menu of pricing plans so that they may select the appropriate plan. Past evidence suggests that this may not be always the case; often customers find themselves in tariffs that do not minimize their costs. An increasingly popular strategy is to recommend plans to customers. In this paper, we examine the effectiveness of this strategy using a randomized field experiment in which some customers were offered plan recommendations and some were not. Our results indicate that encouraging customers to switch to cost-minimizing plans can, surprisingly, harm firm profitability primarily due to an increase in customer churn. We propose two drivers for how the campaign increased churn, namely, by lowering customers’ inertia to switch plans and by enhancing customers’ sensitivity to past overage (i.e., usage beyond the free allowance). Our data provide empirical evidence for both drivers. By leveraging the richness of our field experiment, we assess the impact of targeted encouragement campaigns on customer behavior and firm revenues.

"Idea Generation, Creativity, and Prototypicality”

Marketing

Speaker: Oded Netzer
Associate Professor of Business , Columbia University

18 September 2014 - Room T025 - From 1:00 pm to 2:30 pm


"Idea Generation, Creativity, and Prototypicality”


Oded Netzer (with Olivier Toubia)

Abstract:

In this paper we use text mining and semantic network analysis tools to improve our understanding of the idea generation process. Our contribution is both theoretical and practical. From a theoretical perspective, we link the judged creativity of an idea to the semantic relationships among the set of concepts used to form this idea. The concepts related to any idea generation topic can be represented as a semantic network. Each idea contains a subset of these concepts, which form a semantic subnetwork. The structure of this subnetwork reflects a distribution between novel and familiar combination of concepts. We show a “beauty in averageness” or prototypicality effect, such that ideas with semantic subnetworks that have a more prototypical structure are judged as more creative. We show the robustness of this effect in eight studies across multiple domains. While we focus on judged creativity, the effect also holds with measures of overall idea quality coming from consumers or industry experts. From a practical perspective, we show that our research may be used to automatically text mine and identify promising ideas, and develop online idea generation tools that recommend concepts to users on the fly to help them improve their ideas.


JavaScriptSettings