This article is from the Nonprofit Quarterly’s fall 2017 edition, “The Changing Skyline of U.S. Giving.”

Editors’ note: Fundraising methods have always been sensitive to our communications environment—and right now, with that changing so quickly, individual organizations can find it hard to keep up with the level of testing needed to figure out what will work for them and their donors. And there may be no final there there—since fundraising is also sensitive to issues of oversaturation of any promising tactic we may think up (along with a host of other variables). So maybe it is time to think carefully about how to integrate testing into our fundraising in a more collaborative way, learning how what works for one organization might—or might not—translate to another.

Joint experimentation, knowledge sharing, and collaboration can be a boon to charities that want to maximize their resources to expand their reach and impact. Collaboration between GlobalGiving and DonorsChoose.org has allowed our two organizations to learn more quickly about what inspires donors to give, and enabled us to test joint hypotheses across multiple types of causes and giving platforms. GlobalGiving and DonorsChoose.org have partnered on a handful of tests that coincide with giving days like #GivingTuesday, and we have also looked for the most compelling ways to inspire generosity any day of the year. This article details our experimentation and findings.

GlobalGiving is a global crowdfunding community for nonprofits, donors, and companies in over 165 countries. It helps nonprofits to access the funding, tools, training, and support they need to become more effective. DonorsChoose.org is an education-funding platform that has helped over 350,000 U.S. public school teachers raise over $550 million for classroom supplies, field trips, and guest speakers.

Over the past year, we have secured joint grants from the Bill & Melinda Gates Foundation and the John Templeton Foundation to fund our giving experiments. Until now, most academic research in the nonprofit sector has focused on offline strategies and tactics (like direct mail) that inspire new donors. But these funders have helped us to test academic theories from behavioral economics or social psychology to find out what tactics inspire twenty-first-century generosity online.

The set of answers to that question may sound very familiar. Matching gifts, giveaways, and pseudo-sets can all have a role in helping the public be even more generous. Each of these is explained in its current context below.

Matches, Timing, and Design of the Ask

In the lead-up to #GivingTuesday 2016, the Bill & Melinda Gates Foundation funded a #MatchAMillion campaign on GlobalGiving. GlobalGiving used $500,000 in match funding and A/B testing to find out how it could influence donors to give at higher levels. More than 200,000 donors were presented with a standard matching offer for a one-time donation (“Give now and we’ll match your donation at 50 percent”) versus the option to upgrade to a recurring donation with a higher matching incentive (“If you sign up to make a monthly gift, we’ll match your first month’s donation at 200 percent”). Which offer would drive more giving? GlobalGiving tested the offers via e-mails to its current donors and on the donation checkout page of its website.

Although the standard (50 percent) matching offer via e-mail generated a greater number of one-time donations, the recurring matching “upgrade” incentive on the website generated three times the number of new recurring donations—a statistically significant finding.

These are the takeaways for other nonprofits raising funds online: The data show that nonprofits will be most successful if they keep e-mail appeals simple; appeals should focus on a clear call to action to lead donors out of e-mail and onto the giving page; and the more complex the communications, the harder it typically is to get a donor to act. However, once donors are ready to make a donation, giving them an opportunity to select between two, more complicated matching offers (a lower percentage for a one-time match and a higher percentage for a recurring match) can lead to a significant increase in recurring donations. GlobalGiving has heard from its nonprofit partners that recurring donations are some of their most valuable sources of online funding.

Philanthropic Sweepstakes: Helping to Control the Unknown

In our experience, collaborative experiments among nonprofits can take two forms: trying two distinct approaches that tackle the same problem, or testing the exact same hypothesis in two different environments. In our work with the Gates Foundation on #GivingTuesday, we chose the first option—while GlobalGiving was testing matching campaigns on #GivingTuesday 2016, DonorsChoose.org tested a completely different approach.

Thanks in part to the popularity and success of match campaigns, teachers in record numbers are using DonorsChoose.org to stock their classrooms. As its project inventory grows, DonorsChoose.org has had to develop creative new ways to engage all its donors and teachers when a match offer on all projects is beyond a potential funder’s budget.

One idea continued to surface every few months: a philanthropic sweepstakes. The concept, quickly known around the office as “the golden ticket,” was that donors would be rewarded for certain actions by earning entry into a drawing for a high-value DonorsChoose.org gift card they could then use to support more classroom projects on the site.

#GivingTuesday seemed like the perfect testing ground for the golden ticket. On a day when donors are already primed to give, a small incentive might tilt the scales in favor of classroom projects on DonorsChoose.org.

An added bonus was the chance to promote this #GivingTuesday campaign in advance. Typically, DonorsChoose.org must keep site-wide match days top secret, due to the risk of teachers flooding the site with more projects than the team can review and post while matching funds last. For a typical site-wide match campaign, DonorsChoose.org can give donors and teachers a heads-up twenty-four hours in advance. With the sweepstakes concept, which would involve a set number of prizes but limitless capacity for donations, DonorsChoose.org could promote the campaign weeks in advance without the risk of running out of matching funds.

DonorsChoose.org turned the concept into its first-ever #GivingTuesday GIVEaway. On that one day, every time a donor gave to a project, both the donor and the teacher they supported were entered into a drawing for a $5,000 DonorsChoose.org gift card. Backed by a grant from the Gates Foundation, fifty donors and fifty teachers were declared winners.

Both donors and teachers jumped at the chance to participate, and the campaign raised $1.3 million from 17,217 donors in one day—a year-over-year increase of 155 percent and 184 percent, respectively. November 29, 2016, was a record-shattering day for DonorsChoose.org (until a site-wide match campaign in March 2017 raised over $2 million in twenty-four hours).

The one #GivingTuesday metric that dropped was average donation size, falling from $64 in 2015 to $46 in 2016. The giveaway, which required no minimum donation and limited entries to a maximum of ten per person, indirectly incentivized multiple low-dollar donations to earn extra drawing entries.

A survey among the winning donors revealed that a majority had already planned to make a donation to DonorsChoose.org that day, but a third of survey respondents were enticed by the giveaway concept (10 percent) or by a teacher who asked for their support (19 percent).

DonorsChoose.org also tested various e-mail calls-to-action for donors. Five different e-mails went out to donors on #GivingTuesday: A control that did not mention the giveaway at all; a version that prominently called out the giveaway; a version that only briefly mentioned the giveaway in a “P.S.”; a version that noted that there would be one hundred winners; and a “social proof” version that profiled a donor who had previously won a gift card. Of all the tests, the version that noted there would be one hundred winners converted best.

Our takeaways for the greater online fundraising community: the philanthropic giveaway can be a great tool when a matching campaign is beyond the budget—or on days like #GivingTuesday, when people are already primed to give and see the giveaway as a value add. The response from the winners was overwhelmingly positive (as expected), which can be a great relationship-building opportunity if you can have multiple winners. Still, the match offers remain DonorsChoose.org’s number-one tool for activating its entire community, and match offers work any day of the year. For example, DonorsChoose.org’s #GivingTuesday e-mails from 2015, which mentioned a match, had double the conversion rate of its #GivingTuesday 2016 e-mails about the giveaway. While DonorsChoose.org hasn’t yet replicated the giveaway concept outside of #GivingTuesday, the expectation is that it likely would not exceed the success of the match on a typical day.

Pseudo-Sets: The Missing Piece?

Our collaboration went beyond data sharing in 2016, when the John Templeton Foundation funded a partnership between our organizations and global lending platform Kiva, along with Michael I. Norton and Oliver Hauser at the Harvard Business School, to conduct a large-scale synchronized field experiment.

The purpose of this experiment was to explore how fundraising appeals could be structured in a way to engage donors (or lenders, in Kiva’s case) and increase contributions. Previous research by the Harvard Business School team suggested that people would be motivated to complete tasks if the tasks were framed as part of a “pseudo-set”—pictured as wedges of a pie chart that fill in with each task completed, for example. Inspired by this idea, we ran a large-scale field experiment across the three nonprofit crowdfunding platforms to test the effect of “pseudo-set” framing in fundraising. Could we increase donations by framing an appeal as part of a larger set of tasks?1

All three organizations—GlobalGiving, DonorsChoose.org, and Kiva—launched simultaneous A/B tests on the same day, reaching more than 230,000 past donors to test “pseudo-set” framing. Donors in the test groups received an e-mail that informed them that their past donation was part of a larger set. For GlobalGiving and Kiva, the pseudo-set consisted of the six continents. On DonorsChoose.org, the pseudo-set was made up of six fundamental school subject areas. The e-mails invited people to give (or lend) to a different component of the larger set; completing a pseudo-set meant giving to all six components that made up that set. We found a significant effect (p = 0.084) on donations from pseudo-set framing by two of the charities: GlobalGiving and DonorsChoose.org. We did not find a similar effect vis-à-vis lenders on Kiva. We don’t know why this is so, but we hope to find out in future experiments.

What have we learned from pseudo-set framing of charitable donation requests? It seems that framing donation appeals as part of a larger set is helpful for some organizations. We think that the results can be more powerful if the context and the pseudo-set are presented well. The data from this experiment show that pseudo-sets (the more creative, the better) have the potential to make the donation experience more engaging for donors and help organizations retain their donor base. These are hypotheses that we will continue to test.

Working and Growing Together

Over the past five years, the collaboration between GlobalGiving and DonorsChoose.org has moved from sharing hypotheses to sharing results, and, finally, working together on experiment design, execution, and funding proposals. This would not have been possible without generous funding from the Bill & Melinda Gates Foundation, the John Templeton Foundation, and passionate researchers like Norton and Hauser. And we see great potential for more organizations to work together to uncover other trends and strategies in philanthropic behavior.

If you work with a foundation or in academia, how can you help foster more collaboration in the nonprofit sector with your funding or expertise? Which organizations could you reframe in your mind as potential collaborators instead of competitors? Here are three questions to ask yourself when thinking about collaboration: What other organizations ultimately share your same vision or mission, even if they differ in program or approach? Alternatively, what other organizations employ similar tools or approaches, even if they are looking to impact different populations from those you are looking to impact? And finally, what resources do you have to share (in terms of data, experience, tools, and access) and what resources could you benefit from?

If you identify a potential collaborative partner, the Posner Center for International Development (an international collaborative development community based in Denver, Colorado) has developed a “Collaboration Assessment Tool” that may help you to determine the appropriate level of collaboration and your collaboration readiness.2

It’s not always easy to sell the idea of collaboration to a board or to staff members, who may well be used to working tirelessly to make sure that your organization gets the biggest piece of the pie. But we have experienced firsthand the benefits of using collaboration to grow the whole philanthropic pie instead of competing for crumbs.


  1. Portions of this section on pseudo-sets were taken directly from Oliver Hauser, How “Pseudo-Sets” Might Help Your Nonprofit Get Repeat Donations (Washington, DC: GlobalGiving, 2017).
  2. Available at posnercenter.org/resources/collaboration-assessment-tool.