Michigan State University has 27 fundraising units, each with their own constituency of alumni, donors, and portfolio prospects. In their last capital campaign, from July 2011 to December 2018, MSU raised $1.83 billion from 255,284 donors. The goal of this project was to develop predictive models for each of Michigan State University’s 27 fundraising units to identify living individual donors who were likely to make a philanthropic commitment of any amount in the next 12 months. Identifying likely donors to each of the fundraising units would allow for increasingly strategic solicitation of a unit’s constituents. A review of the literature examined philanthropic giving models from other institutions to inform the methodology used in this study. Time-stratified model evaluation was used to examine 16,200 different models and select the best combination of data resampling technique, tree-based classification method, and hyperparameters for each fundraising unit. Among the 27 best models, the accuracy ranged from 83% to 99%. Global and individual-level interpretations for the models were developed using SHAP plots to provide information on how the input predictor variables influenced the predicted likelihood to increase buy-in from end users. The findings will be used to engage and solicit fundraising unit constituencies in an increasingly data-driven way.