18 May Understanding Why Donors Do (Not) Return
Guest post: Tim Althoff is a PhD candidate and Jure Leskovec is a professor in the Computer Science Department at Stanford University. Both are interested in research on large-scale traces of human activity to better understand human behavior.
Recruitment and continued engagement of donors is critical for the success of crowdfunding communities, such as DonorsChoose.org. With donor attrition rates above 70%, a significant challenge for online crowdfunding platforms as well as traditional offline non-profit organizations is the problem of donor retention — getting donors to continue to make donations year after year. Increasing donor retention would have significant impact on the effectiveness of fundraising campaigns, as even small improvements in donor retention can have a significant impact on the amount of collected funds. For example, a 10% improvement in donor retention could yield up to a 200% increase in obtained donations.
Despite the importance of donor retention for fundraising campaigns, many of its basic aspects are still not well understood. Current knowledge about donor retention largely consists of anecdotal evidence from fundraising professionals and small lab experiments in artificial environments. There are many questions about donor retention that remain open. For instance, are different donor subgroups affected differently by timely acknowledgments? What does “timely” even mean and what can we infer about the donor’s expectations from their behavior?
Donor Retention on DonorsChoose.org
Through DonorsChoose.org‘s data, we are able to empirically study donor retention behavior across millions of donors and donations. This allows for an unprecedented and detailed view of how people direct their donations. We explore various factors impacting donor retention, specifically for first-time donors, which allows us to identify different groups of donors and quantify their propensity to return for subsequent donations.
Our results are described in this paper published this week at the International World Wide Web (WWW) Conference. We found that only 26% of first-time donors ever return and donate a second time. However, it is important whether or not a donor’s first project succeeds and gets fully funded. Donors that experience early success are 29% more likely to return compared to those whose first project fails. This finding has led to improvements in project exploration for first-time donors on DonorsChoose.org.
The Donor’s Role Within the Project
We also found that donors assume different roles within a project (see plot below). There are starters that like to start off new projects with an initial donation, closers that like to finish off projects that are close to completion, and a third group that does not particularly follow any of the previous two behaviors. We find that closers and starters are more likely to return than other donors. Starters tend to be teacher-referred donors that live in close proximity to the project they are supporting. Closers tend to be site donors (not teacher-referred) that often live further away. Both donors that live very close (within 25km) or very far away (farther than 2000km) are more likely to return, which partially explains the “U-shape” in the plot.
Timely Communicating Impact to Donors
Communicating impact to donors has been identified as an important driver of donor loyalty. After a project gets fully funded on DonorsChoose.org, teachers are asked to write a letter that expresses the impact that the donated materials will make on their classroom. We measured how donor return rates varies with different teacher response times (number of days since the project was fully funded; see plot below). Donors that receive an impact letter within the first week or month are much more likely to donate to the same teacher again compared to donors who have to wait several months or maybe never receive such a letter (“NA”).
Predicting Donor Return
We found that a variety of other factors such as entering the community through different means, geographical patterns of donations, the donation amount, and disclosure of optional personal information all shine light on the donor’s initial motivations and commitment. Furthermore, factors including project cost, project success, and timely responses by the teacher (highlighting the difference the donor has made) can affect the donor’s sense of personal impact and trust in the organization. In addition, we showed that the teacher’s ability to retain donors is correlated with their experience, timely writing thank you notes, and use of Facebook for solicitation.
Based on these insights, we were able to predict whether a donor will return for a second donation based on just the properties of the donor’s first donation with promising accuracy. The model reveals that there are some donors that are very unlikely to return while others are more likely. Focusing on the latter group in fundraising campaigns could lead to higher efficiency and significant savings.
Many more donor retention factors and discussion can be found in the full paper of our study.