We review panel data models popular in marketing applications and highlight some issues, potential solutions, and trade-offs that arise in their estimation. Panel data studies controlling for unobservables often show dramatically different estimates than cross-sectional studies. We focus on models with unobservable individual-specific effects and address some misconceptions appearing in marketing applications.
Browse by title
Natalie Mizik and Eugene Pavlov
Natalie Mizik and Robert Jacobson
We illustrate the application of dynamic panel data methods using the direct-to-physician (DTP) pharmaceutical promotions data described in an article by Mizik and Jacobson (2004). Specifically, we focus on using panel data methods to determine appropriate model specification and to demonstrate how dramatically the estimates of the DTP effectiveness change across various common model (mis)-specifications.
Rahul Guha, Darius Onul and Sally Woodhouse
We outline some basic considerations and implementation strategies regarding the use of consumer surveys and conjoint analysis in the context of complex litigation. We also describe two applications of these techniques in antitrust disputes in the payment card and infant formula supplements industries.
Online advertising has grown rapidly in recent years. The rise of this new form of advertising has generated a number of policy questions around privacy, the ability of local governments to regulate information, and antitrust in online markets. This chapter reviews three studies using a combination of field experiments and quasi-experimental variation to answer policy questions related to online advertising.
Pradeep K. Chintagunta
In this chapter, I provide brief discussions of what we mean by structural models, why we need them, the typical classes of structural models that we see being used by marketers these days, along with some examples of these models. I provide a basic discussion of structural models in the context of the marketing literature and limit myself largely to models of demand rather than models of firm behavior.
Alan G. White and Rene Befurt
We discuss the use of regression analysis to evaluate harm in a breach of contract case involving allegations that the licensor of a product failed to use commercially reasonable efforts to promote and sell the product. Regression analysis has been widely used and accepted by US courts across a large variety of different types of cases, including labor discrimination cases, antitrust cases, and intellectual property cases. In cases involving marketing issues, regression analysis is frequently used to determine the effect of promotion on sales.
Marnik G. Dekimpe and Dominique M. Hanssens
Determining the long-term impact of marketing actions is strategically important, yet more challenging than uncovering short-term results. This chapter describes persistence modeling on time-series data as a promising method for long-term impact detection, especially as longitudinal databases in marketing are becoming more prevalent. We provide a brief technical introduction to each step in persistence modeling, along with a set of illustrative marketing studies that have used such models. Next, we summarize various marketing insights that have been derived from the use of persistence models in marketing.
Michael Trusov and Liye Ma
Constructing behavioral profiles from consumer online browsing activities is challenging: first, individual consumer-level records are massive and call for scalable high performance processing algorithms; second, advertising networks only observe consumer’s browsing activities on the sites participating in the network, potentially missing site categories not covered by the network. The latter issue can lead to a biased view of the consumer’s profile and to suboptimal advertising targeting. We present a method that augments individual-level ad network data with anonymized third-party data to improve consumer profile recovery and correct for potential biases. The approach is scalable and easily parallelized, improving almost linearly in the number of CPUs. Using economic simulation, we illustrate the potential gains the proposed model may offer to a firm when used in individual-level targeting of display ads.
Daniel M. Ringel and Bernd Skiera
Visualizing competitive relationships in large markets (i.e., markets containing over 1,000 products) is challenging. We discuss a new model called DRMABS (Decomposition and Re-assembly of MArkets By Segmentation) for such applications. DRMABS combines methods from multiple research disciplines such as biology, physics, computer science, and sociology with a new method of submarket-centric mapping to visualize asymmetric competition in large markets in a single two-dimensional map.
This chapter addresses some of the practical implications involved in conducting case study research including finding a suitable research question, negotiating access, and gaining participation. The chapter also delves into the less spoken of area of interference in the research while finishing with a discussion on the project management aspect of case research to ensure the research is kept on track. Each section provides a ‘lessons learned’ section providing practical advice for case researchers.