The authors address the situation in which a researcher wants to cross-tabulate two sets of discrete variables collected in independent samples, but a subset of the variables is common to both samples. The authors propose a statistical data-fusion model that allows for statistical tests of association using multiple imputations. The authors illustrate this approach with an application in which they compare the cross-tabulation results from fused data with those obtained from complete data. Their approach is also compared to the traditional hot-deck procedure.

A variant of Banker and Morey's (Banker, R. D., R. C. Morey. 1986. The use of categorical variables in data envelopment analysis. Management Sci. 32(12, December) 1613–1627.) DEA model for controllable ordinal outputs. As opposed to the original model, this version allows the comparison of a Decision Making Unit (DMU) to other DMU's operating at equal or higher levels of the ordinal outputs.

This article provides a descriptive analysis of how methodological factors contribute to the accuracy of customer churn predictive models. The study is based on a tournament in which both academics and practitioners downloaded data from a publicly available Web site, estimated a model, and made predictions on two validation databases. The results suggest several important findings. First, methods do matter. The differences observed in predictive accuracy across submissions could change the profitability of a churn management campaign by hundreds of thousands of dollars. Second, models have staying power. They suffer very little decrease in performance if they are used to predict churn for a database compiled three months after the calibration data. Third, researchers use a variety of modeling “approaches,” characterized by variables such as estimation technique, variable selection procedure, number of variables included, and time allocated to steps in the model-building process. The authors find important differences in performance among these approaches and discuss implications for both researchers and practitioners.

In a situation where several hundred new music albums are released each month, producing sales forecasts in a reliable and consistent manner is a rather difficult and cumbersome task. The purpose of this study is to obtain sales forecasts for a new album before it is introduced. We develop a hierarchical Bayesian model based on a logistic diffusion process. It allows for the generalization of various adoption patterns out of discrete data and can be applied in a situation where the eventual number of adopters is unknown. Using sales of previous albums along with information known prior to the launch of a new album, the model constructs informed priors, yielding prelaunch sales forecasts, which are out-of-sample predictions. In the context of new product forecasting before introduction, the information we have is limited to the relevant background characteristics of a new album. Knowing only the general attributes of a new album, the meta-analytic approach proposed here provides an informed prior on the dynamics of duration, the effects of marketing variables, and the unknown market potential. As new data become available, weekly sales forecasts and market size (number of eventual adopters) are revised and updated. We illustrate our approach using weekly sales data of albums that appeared inBillboard'sTop 200 albums chart from January 1994 to December 1995.

With recent advances in information technology, most companies are amassing extensive customer databases. The wealth of information in these databases can be useful in identifying those customers most likely to purchase a new product and in predicting when this adoption may take place. This can assist database marketers in determining when individuals should be targeted for the promotion of a new product, which may increase the efficiency of manufacturing and distribution, and assure a faster return on investments. For this purpose, we propose a model that considers the timing of past purchases across multiple product categories and produces estimates of each customer's propensity of ever purchasing in a particular product category and of the timing of their purchases. The model is designed to help managers identify the best prospects for a new offer in one of multiple categories based on generalizations obtained from past offers. The proposed model also provides projections of aggregate penetration for new brands within the database, based on sample estimates.