Database Paper—The IRI Marketing Data Set

Abstract

This paper describes a new data set available to academic researchers (at the following website: http://mktsci.pubs.informs.org ). These data are comprised of store sales and consumer panel data for 30 product categories. The store sales data contain 5 years of product sales, pricing, and promotion data for all items sold in 47 U.S. markets. In two U.S. markets, the store level data are supplemented with panel-level purchase data and cover the entire population of stores. Further information is available regarding store characteristics in these markets. We address several potential applications of these data, as well as the access protocol. The data set described in this paper is maintained by IRI. Any fees charged by IRI for the distribution of the data set will be used for the continual maintenance and updating of the data. Scholarships to cover IRI's fees (for those who need it) are available through the INFORMS Society for Marketing Science (ISMS). Please see the website above for further details.

Consumer Stockpiling and Competitive Promotional Strategies

“…Although Rao et al find that prices across brands are independent, they note that in many cases distributions have more than two modes. Our examination of the distribution of prices of multiple stock-keeping units (SKUs) across several categories in the IRI marketing data set (Bronnenberg et al 2008) is consistent with the authors' finding that the distribution of prices is often multimodal and that prices frequently exhibit negative serial correlation.…”

“…that the distribution of prices is multimodal. A detailed description of the data is available in Bronnenberg et al (2008). The evidence presented here is from the 2006 in the data set.…”

Editorial—Maximizing Impact via Database Submissions

“…Our hope at Marketing Science is that the IRI marketing database paper (Bronnenberg, Kruger, and Mela, 2008), published here, can have similar impact. So, how does it count?…”

“…As Marketing Science publishes its first database paper in this issue (Bronnenberg, Kruger, and Mela 2008), has two more submitted in the pipeline, is receiving a large number of current submissions on Marketing citation analyses to assess impact (a good thing, and reflecting increased interest), and I receive (legitimate) questions such as "What is a database paper worth on one's vita?" (which I will answer below), it seems like an opportunity to highlight the issue of maximizing impact, "broadly defined."…”

A Model of Heterogeneous Multicategory Choice for Market Basket Analysis

“…For our binary data, this corresponds to the absolute deviation. We use the IRI data set explained in detail by Bronnenberg et al (2008) and Kruger and Pagni (2008) to test our model. The IRI data set contains information which product categories are bought by each household in several outlets.…”