Analysis of factors influencing recurring donations in a university setting: A compound poisson mixture regression model

Abstract

We present a Compound Poisson Mixture Regression model of the joint distribution of transaction frequency and monetary value, and apply it to study alumni donations at a university in the USA. The model captures covariate effects, recognizing that both response variables emanate from one statistical unit — a donor. Heterogeneity, group-level factors, and other features of the data are captured through coefficients that vary between segments.The data in the study are transaction records for the 2000–2016 period, and a survey conducted in 2017. Despite including subjective factors from the survey, the results suggest that between-segment differences are unobserved. Heterogeneity is manifested in covariates, including subjective factors – psychological distance, perceptions of donation impact, willingness to volunteer – displaying stratified effects on either transaction amounts, frequencies, or compound effects on both variables. Characterization of such effects supports the development of tailored fundraising/marketing strategies aimed at increasing donor retention and lifetime value.