Â鶹ÒùÔº

Nonprofits can use public data to target fundraising appeals by state and ZIP code

While nonprofit organizations are growing, their donations are shrinking. The number of nonprofits registered with the Internal Revenue Service grew 25% from 2013 to 2023. But during the past year, both money raised and donor counts have dropped 3%.

Their main challenge is low response rates for fundraising solicitations, says Vijay Mahajan, professor of marketing at Texas McCombs. In turn, a major reason is lack of quality data on donors, which makes it harder to successfully target their appeals.

Nonprofits tend to keep data on active donors, such as how much and how often they donate. But it's often too expensive to gather and maintain data on others, such as prospective donors.

In a study in the Journal of Business Research, Mahajan identifies a workaround. Instead of using actual donation histories to predict who might respond to fundraising appeals, nonprofits can use community-clustered profiles.

The profiles use readily available public data to identify where potential supporters are more likely to live, based on demographic, financial, and social characteristics.

"Nonprofits can learn a lot about their donors using the data in community profiles," Mahajan says. "If you go to secondary sources for data at the state and ZIP code levels, you can tell a lot in terms of who is going to respond to you."

Picturing potential donors

Typically, Mahajan says, fundraising teams appeal to three types of donors. They try to keep active ones while reactivating lapsed ones and acquiring new ones. Community-clustered data, he believes, might help them target all three.

His underlying theory is that birds of a feather flock together. People with similar backgrounds and lifestyles tend to gravitate into the same communities.

To test the concept, he partnered with Shameek Sinha of the University of Auckland, Sumit Malik of the University of Liverpool, and the late Frenkel ter Hofstede of Texas McCombs. They gathered publicly available survey data by state and ZIP code.

The largest dataset came from the Direct Marketing Educational Foundation, which sends direct mail solicitations to individuals who have donated at least once. It covered 429,310 donors over 15 years. Other data sources covered communities by lifestyle, crime rates, and religion.

The researchers' resulting model mapped 44 characteristics to create profiles for each community. Gender, household size, and financial status had the most influence on the potential to donate.

Based on those profiles, they predicted the success rates for targeting appeals for each donor group. When they compared their predictions with actual donation behavior, the success rates were similar.

That means that when fundraisers are short on donor information, they can use such profiles instead, Mahajan says: "You don't have to know all the details about your contributors. Just find out where your people live."

Practical uses for profiles

Nonprofit managers can use the study's findings in several ways, the researchers say.

• Mix and match. Fundraising teams can add community-clustered profiles to their existing data on donors to improve response rates.
• Seek potential donors. Where information on possible new donors is sparse or costly to acquire, nonprofits can use profiles to identify the communities in which they're most likely to live. "They identify donors in that community who match that profile and target them with their donation solicitation efforts," Sinha says. "This way, nonprofits can avoid spending millions of dollars to collect and use behavioral data."
• Awaken dormant supporters. Community profiles were most accurate at identifying lapsed former donors who could be persuaded to contribute again.

Best of all, for resource-strapped nonprofits, the data is there for the taking, Mahajan adds. "You don't have to conduct the research on your own. It's already been collected. Just that information itself can tell you a lot about the people who are contributing or not contributing."