The application of business analytics has been widely explored in for-profit contexts; however, its adoption within nonprofit organizations, particularly in fundraising decision-making remains limited. This study examines how predictive analytics can support strategic fundraising by anticipating donor behavior in non-governmental organizations (NGOs). Drawing on a systematic literature review, the research adopts the CRISP-DM methodology to develop and compare five predictive classification models using historical donor data from a Portuguese NGO. The models are evaluated using multiple performance metrics, including accuracy, specificity, sensitivity, F1-score, and area under the ROC curve. Results indicate that a CHAID-based model achieves the strongest overall performance, enabling the identification of donor profiles with high and low propensities to give using variables commonly available in nonprofit databases. In parallel, an interactive analytics dashboard is developed to translate model outputs into actionable managerial insights, supporting fundraising planning and resource allocation. The study highlights both the strategic value of analytics-driven fundraising and the practical challenges associated with data quality, technical expertise, and organizational readiness. By demonstrating how business analytics can be operationalized in a nonprofit setting, this research contributes to the growing literature on analytics-enabled decision-making beyond the for-profit domain.

--

Business analytics,Predictive analytics,Fundraising strategy,Donor behavior,Nonprofit organizations

• Computer and Information Sciences - Natural Sciences
• Economics and Business - Social Sciences