The authors develop a Bayesian approach for inferring the joint distribution of several demographic variables when only information about the marginal distribution of each variable is known and when there is prior information about the correlation between the variables. Using public domain U.S. Census Bureau geodemographic data for Sioux Falls and the state of South Dakota, they apply the approach to four specific marketing problems, two involving direct mail advertising campaigns and two involving the location of a retail site. The empirical applications illustrate the serious bias that results from the assumption that the demographic variables are independent of each other. Overall, the Bayesian approach recovers the underlying joint distribution of the variables of interest well and is unique in its ability to provide a measure of uncertainty about the resulting estimates. Hypothesis testing and the construction of confidence intervals are illustrated in the context of a retail site location application. The results of a Monte Carlo experiment indicate that the Bayesian approach recovers joint densities under a variety of conditions in a robust manner.