What if it were possible to predict who was likely to donate to charity — before they ever gave any money?
Researchers from the University of Texas at Austin have found a way to do just that.
Shameek Sinha, PhD 2011, and McCombs School of Business Professors Frenkel Ter Hofstede and Vijay Mahajan created a model that helps predict whether someone who has never before donated to a specific philanthropic organization is likely to give money to that group, as well as predicting the size and frequency of those donations.
Those predictions are based on data about the potential donor's community, such as the religion, ethnicity, and political affiliation of people in that area. Although past donation behavior is typically considered the best method for predicting people’s future giving, when that information isn’t available, the researchers’ findings suggest community data can be used instead.
For the particular philanthropic group the researchers considered, certain communities proved more likely to give. “If you want people to donate, you could target someone who is a white Republican” said Ter Hofstede, associate professor of marketing.
Their working paper, tentatively titled "Duration Dependence of Heterogeneous Donation Behavior: How Credibly Can We Predict Behavior Using Community Characteristics of Donors?" stems from Sinha’s PhD dissertation.
This upcoming paper, which was co-authored by the two professors, grew out of Ter Hofstede’s earlier research into consumer response to catalog mailings. The philanthropy study applied Ter Hofstede’s earlier approach to a different context: Creating a model to help charitable groups increase new donations by mailing more carefully targeted solicitations to existing and potential donors.
To do that, the researchers looked at data from the nonprofit Direct Marketing Educational Foundation (DMEF), which revealed where donors lived, the timing of both solicitations they received and contributions they made, and how much money donors gave to a specific nonprofit. (The name of that philanthropic organization remained confidential, however, even to the researchers themselves.) Alongside data on neighborhood demographics and finances, the researchers also looked at county level information on religious beliefs, political affiliations, and crime rates in specific zip codes across Texas.
The Ideal New Donor
Who would make an ideal donor? It varies, of course, based on the charitable organization’s mission. But using the data associated with the specific philanthropic organization they considered, the researchers said an ideal donor — one who contributes both more frequently and larger dollar amounts when they donate — fit the following profile:
- White Republican
- Member of a smaller-family household
- Younger, more educated
- Has a higher income, wealth rating, and higher home value
- Carries older trade-lines (lines of credit, such as loans and credit cards) with higher balances
- Carries more trade-lines with satisfactory ratings, fewer trade-lines with derogatory or delinquent ratings (late or unpaid debts)
- Rather than being a Catholic is a member of an Evangelical or other Christian denomination
- Live in a neighborhood with a greater concentration of public schools, with better school ratings (a weighted combination of dropout rates and SAT scores)
Applying the Methodology
But what about charitable organizations that are more likely to attract a different type of supporter — say, indebted college students donating to President Obama's campaign? They can still apply the same methodology by plugging data into the researchers’ model. That enables the charity to figure out what type of potential new donor their organization should target with a mailing. “Even if you do not have information on your [potential] donors’ past giving behavior, if you just look at your communities or the communities where these donors live, that does matter” in predicting their donation behavior, Sinha said.
From that starting point, as additional donor information is collected over time, organizations can better fine-tune their solicitation efforts targeting both new and current donors.
“Since the economic downturn in 2008, most nonprofits have struggled to acquire new donors,” said Keary Kinch, McCombs’ director of alumni relations and annual giving, who has 30 years of experience in the direct marketing fundraising business. “This type of targeting could very well supply the additional boost to response rates that can make acquisition a winning proposal again,” Kinch said.
As an example, for the nonprofit studied, directing mailings to a neighborhood made up of “white Republicans or with higher home value or something like that, that’s definitely a better way than just ad hoc targeting,” Sinha said. He explains that with ad hoc targeting, the random nature of solicitations makes it difficult to know how potential donors may respond to your mailings.
“But once you know that a neighborhood with white Republicans gives you more, you obviously put your marketing dollars more there, because that’s where the expectation of returns are at the highest,” Sinha said.
“Communities do matter,” he said. “That’s the basic idea.”