Fundraising campaigns in Northern California have been analyzed using predictive analytics models to determine the degree to which past behavior, demographics, charitable giving, wealth profiles, and neighborhood characteristics predict response to direct mail appeals. In this session, the presenters will compare a half dozen different models that predict which past donors will give to a current campaign and models that predict how new prospects will respond. We will discuss practical approaches for targeting direct mail. Campaigns can be made more cost effective by eliminating mail sent to people who are unlikely to respond and by analyzing variables to assess which characteristics most influence response. Interesting patterns were identified such as the proximity curve — how response varies by distance to the land conservation project, and the donor decay curve — the mathematical function that describes how current donors become lapsed donors. The findings can be used by any land trust and have broad implications for data collection, estimation methods, targeting, revenue optimization, and the procurement of lists for future appeals.