Machine learning isn’t a buzzword, it’s a practical tool advancement teams can use today to improve donor engagement and retention. This session will show how institution-built ML models can predict donor attrition with surprising accuracy and help advancement leaders make smarter, data-informed decisions.

I will highlight a working attrition model built with Stony Brook’s donor history, where advancement experts validated the list of “most at-risk” donors as highly accurate. The lesson is simple: models trained on your own donor data outperform vendor black boxes. By leveraging the patterns in institutional giving history, shops can maximize unique donor counts, focus stewardship efforts, and strengthen retention.

The session will also cover how to prepare donor datasets for modeling, approaches to validating results, and ways to get started with machine learning even without a dedicated data science team. Participants will walk away with a clear understanding of how ML can support fundraising and stewardship when applied directly to their own donor data.