
Custom predictive modeling
Summary
Fundraising requires precision. Whether you are scaling your development team or managing a rising postage budget, you cannot afford to solicit the wrong people.
Our Custom Predictive Modeling offering is tailored specifically to your data. We do the extraction and use the best techniques in machine learning to provide likelihood scores that minimize false positives, reducing costs and yielding more donations.
Features
Spouse-linking
Historical sampling
CRM/marketing activity and fundraising metrics
Dimensionality reduction
Ensemble learning
Experimental LLM and natural language processing
Scoring for major giving, annual giving, planned giving, and more
What to expect
Initial consultation — We briefly meet to determine your scoring needs.
Onboarding — You provide access to your systems.
Discovery — We meet to tailor our models to your interests.
Build — We extract data, perform feature engineering, and then train, test, and validate multiple models to find the best fit.
Delivery — We score your constituents using the best-performing model, present our scores by tiers of likelihood, and can even build them into your CRM to be responsive to updated data.



