Individual donations can be a highly effective way for nonprofit organizations to raise money and ensure their financial health for years to come. The catch is, fickle donation patterns can result in a roller-coaster ride of peaks and valleys.
“If only there were a way to predict donors’ behavior,” you might think. “We’d be able to identify sources of more donations, and our organization would be better able to plan our operations.”
Actually, there is such a method. It’s called donor predictive modeling, and it’s fast becoming a best practice among both large and smaller nonprofit organizations. This method of projecting income and financial health relies on the tried-and-true science of computer modeling, which has been used for several decades by companies in highly technical industries to develop and test new products.
A quick look at predictive modeling
To cite just one longstanding example, aerospace companies use predictive modeling to test how hardware and software on jets, rockets and satellites work together before any of these vehicles leave the ground. This lets engineers solve issues earlier in the development cycle, saving money and boosting vehicle safety.
Similarly, nonprofits can use both simple and more complex predictive models to more accurately pinpoint sources of revenue beyond their base of high-income contributors. Smart nonprofit leaders have long realized that less-affluent people can be just as generous with their money as wealthier people (and sometimes more). And the sheer numbers of potential middle-income donors make them worth pursuing.
Three essential varieties of predictive modeling
In a white paper for the nonprofit consultancy Blackbaud, Lawrence Henze describes three types of predictive models: generic, prescriptive and custom.
- A generic model bases its predictions on “big data,” such as information about prospective donors’ zip codes. The idea is that because residents within one particular zip code can be significantly wealthier than those in a neighboring one, they are more apt to donate because they have more money to give. Generic models also consider variables such as education levels, age and family status. The trouble with generic modeling is that it can’t identify people’s behavioral tendencies, which can strongly predict the propensity to give.
- A prescriptive model takes more personal information into account, such as a person’s history of giving to an organization and to other groups. This helps identify behavioral tendencies, so it can be better than generic modeling, though it’s still not optimal. As the oft-repeated tag on investment ads proclaims, past performance does not predict future results.
- A custom model, developed just for the organization doing the analysis, is the best kind of model. For instance, it can dissect the behavior of both donors and non-donors in an organization’s database: people who have been asked to donate but who have not yet given. It also analyzes the ties each individual stakeholder maintains with your organization. Besides donation histories, such a model includes information on volunteer efforts, board service and attendance at major fundraising events. Using a custom model, an organization can craft campaigns that aim different messages to people in a multitude of cohorts. The organization can even customize an individual appeal letter or email that speaks to each recipient’s background, motives, philosophy, financial situation and other compelling factors.
Success stories abound
Donor predictive modeling is a worthwhile endeavor for any nonprofit for the simple fact that it works well and it’s relatively inexpensive to implement. In New York City, for example, the venerable Whitney Museum of American Art used predictive modeling to boost fundraising for a new museum building.
Working with the consulting organization Rapid Insight, the museum’s development team came up with a computer model that could find patterns in the membership and giving histories among affluent people who were nonetheless not contributing as generously as other established donors in the museum’s database. This insight helped the museum find ways to increase donations and identify new donor prospects.
The Whitney team has continued to alter its model. It now can target donors in tandem with the museum’s acquisitions and exhibit schedule, for instance. In addition, it is exploring ways to mimic donor campaigns at colleges and universities, which have been standard-bearers for donor retention.
Most organizations will need to hire a consultant to perform such exercises, especially when just starting out. Yet, ultimately, this will be the kind of work that any group’s IT staff should be able to perform handily.
Growth in giving
Some nonprofits are thinking on a much grander scale. The Growth in Giving (GiG) database, which began in 2012, now contains more than 100 million transactions, including data such as date and donation amount. The names of individuals and organizations are encrypted, but the GiG database can nonetheless track behavior patterns over time.
While the GiG database points to the need for nonprofits to do a better job of donor retention, it also identifies one somewhat magical number: $250. According to its data, once a person contributes this amount to an organization, he or she is likely to remain a reliable donor in the future.
Here, too, the old marketing adage holds true. Just as it is considerably more expensive for a for-profit company to find a new customer than to retain an existing one, donors are also more expensive to find than they are to retain. Donor predictive modeling can thus be an extremely valuable tool in conserving and building upon a valuable resource: a nonprofit’s current crop of donors.
Tags: Business LeadershipLearn More: Click to view related resources.
- Lawrence Henze, "Using Statistical Modeling to Increase Donations," Blackbaud.com
- "Why Nonprofits Should Be Building Predictive Models," Rapid Insight Inc.
- "Growth In Giving Database Offers Behavior Analysis," Nonprofit Times
- Adrian Sargeant, "Donor Retention: What Do We Know & What Can We Do about It?," Nonprofit Quarterly