Acknowledgments
This work would not have been possible without a generous grant from the Bill & Melinda Gates Foundation. We would like to extend special thanks to Akruti Desai for her continued advice and support in building and testing this giving infrastructure, as well as the ideas42 and the Better Giving Studio teams for their prior research into and development of their original GiveLists concept. We also would like to thank our team members who worked to bring this project to life, including Mark Ulrich, Franceska Rolda, Rahul Gupta-Iwasaki, Tina Roh, and P22 Studio.
Introduction
Choice overload is a common barrier to charitable giving, even when a donor knows what cause they want to support. With over one million charities in the United States, the overwhelming number of choices can cause people not to donate at all. In order to address this problem and ease the burden of choice for donors, the Better Giving Studio proposed the concept of GiveLists. Prior research by ideas42 has demonstrated the effectiveness of expert curation in increasing donations and found generic curation—lists made by peers—could potentially be just as effective in inspiring greater giving.
One of the pitfalls of expert curation is that identifying and recruiting experts is time and resource intensive, prone to bias and resulting in only a small number of voices being heard. While expert curation encouraged a subset of donors to give more, some donors were “reluctant to engage with expertly curated lists for one reason or another”. Ideas42 researchers concluded, “Dynamic lists that incorporate the latest information or respond to current events could be a better tool for conscientious donors. Similarly, personalized or localized lists could offer more salient recommendations for those who seek bespoke guidance.” Furthermore, they proposed that presenting a simplifying option to donate to all nonprofits on a list increased the likelihood to give.
As a nonprofit building an accessible giving infrastructure, Every.org wanted to explore these concepts further and help bring this concept to life, with a focus on scalability and ability to support any curation method. With the support of the Bill & Melinda Gates Foundation, Every.org set out to build out an infrastructure flexible enough to support any methodology for curation, up to any number of lists.
By utilizing Every.org’s existing recommendation engine and backend, in under 7 months we built out a public API to host and serve lists of nonprofits, launched a fully functional website implementing the option to split a donation to all nonprofits on a list open-sourced to the public, and started testing how different methods of curations impacted donors making real donations to nonprofits.
The default behavior of Every.org lists was that the donation was split equally among all of the nonprofits, but a donor could still explore the individual nonprofits on the list and support them separately.
While it was a short period, we launched 27 lists that collected over 1000 donations for charities, with a cumulative total of nearly $86k. Below is a summary of our takeaways, which we expand upon in the following sections.
- Lists are primarily used for temporary fundraising.
- Lists can succeed even with a mix of cause areas.
- Lists enable collaborative fundraising, but nonprofits are short on resources.
- Dynamic lists provide a scalable path forward.
Lists are primarily used for temporary fundraising.
The biggest takeaway was that curators of giving lists were not interested in creating permanent (or semi-permanent) lists and instead wanted to use them for time-bound fundraising efforts. We reached out to experts, nonprofits, and communities about their interest in curating lists of nonprofits. Through the website, individuals were also able to create their own list by filling out a survey form.
The vast majority of the created lists were intended to be temporary fundraisers. These included a holiday giving list, a charity cycling event, a community launch event, and an abortion fund list in response to news about the state of Texas’ restrictions. Donations to lists tended to peak during specific time periods ranging from a day to a month, and then tapered off after the event had passed.
On the other hand, we had difficulty generating excitement for more persistent, static lists. We launched a site featuring 10 curated lists based on cause areas including “Education Everywhere”, “Tackling Climate Change”, “Ending Global Poverty”. The nonprofits on these lists were based on partner suggestions and our own research. We promoted across social media channels and put matching funds to incentivize giving, but ultimately each list received less than 10 donations— much fewer in comparison to lists created for time-bound fundraisers. While there are many factors that could have contributed to the lack of success, there was generally much less enthusiasm from curators and fewer donations overall to these static giving lists.
We eventually had to fold this site into our main donation platform due to trademark issues, and due to the lackluster response overall, we do not plan on supporting this site long term. Lists will still be supported through our API and the source code for the original site has been made open-source.
Overall, lists were primarily used as a fundraising tool to convert excitement into action during a time-bound event.
Lists can succeed even with a mix of cause areas.
Previous research featured lists of nonprofits that were all in the same related cause area due to the focus on expert curation. Expertise usually implies that the curator focuses on a limited number of nonprofit cause areas, such as “animals” or “education”. We found that a list can encourage giving even if it contains a mix of cause areas that are not directly related to each other.
The most successful case was the Better World Fundraiser created by Lil BUB’s Big Fund, a nonprofit organization focused on raising funds to support animals who are special needs. Their team created a Better World Fundraiser promoting a list of nonprofits from a variety of causes: Sycamore Land Trust (environment), Power of Zero (prevention of youth bullying), and Milo’s Sanctuary (cat sanctuary), and their own nonprofit. They rallied supporters on various social media channels with a message to create a better world, resulting in 598 supporters giving 659 donations totalling over $56k and including several monthly recurring donations.
Lists enable collaborative fundraising, but nonprofits are short on resources.
Just as it is difficult to find and recruit expert curators, it was time and resource intensive to inspire communities and nonprofits to band together to fundraise using lists. Given that nonprofits are experts in their own fields, we were wondering if organizations with complementary missions would be interested in curating a list of organizations working on complementary missions together.
The main selling point was that donors could elect to share their contact information with all of the nonprofits on a list, and fundraising together would allow nonprofits to share donors with whom they could build lasting relationships with.
The reaction was mixed, with reluctant nonprofits concerned about fulfilling their own fundraising needs and feeling uncertain about how their donors would react to such an effort. Even in cases where nonprofits were interested in collaborative fundraising—typically in the context of a giving or awareness day—many were simply unable to commit the time and resources to coordinate such an effort.
The nonprofits that created lists were either already focused on fundraising for other organizations or already had very close relationships with them. This made activation around collaborative fundraising easy and low lift.
Still, lists required active promotion and influence to succeed. Lists tended to receive the most donations during times of active promotion, and donations tapered off a few days after the final promotion.
In the successful case of the Better World Fundraiser, the list was created by a grantmaking organization who led promotion without requiring too much fundraising effort from the other supported nonprofits. The smaller organizations on the list benefited from Lil BUB’s wide audience, including over 800 thousand followers on Twitter and 2.4 million on Instagram. The Lil BUB team highlighted the list frequently in their social media accounts and included it in their newsletter. Along with consistent promotion, they also had live Instagram interviews with each of the organizations highlighted on the list.
Ultimately, we were largely unsuccessful in convincing nonprofits to collaboratively fundraise with each other within the short timeframe available, given that they are already strapped on time and resources. There seems to be some potential in using lists as a tool for time-bound collaborative fundraisers with sufficient planning and incentives.
Dynamic lists provide a scalable path forward.
While there is promise in lists being used for fundraising purposes, we were curious whether there was a method of curation that could be scaled and remain useful for guiding giving decisions outside of normal fundraising. Curated lists of nonprofits all face the same issue: who will curate these lists? There are hundreds of cause areas in the charitable sector. Each donor may have different preferences among countless other dimensions, and it is impossible to manually find curators for each one of these permutations. A potential solution to this issue could be algorithmic recommendations driven by machine learning.
Artificial intelligence can generate lists instantaneously on a massive scale and even customize for each donor. Similar to recommendation engines on platforms like Pinterest, we can help individuals find nonprofits in cause areas they are interested in through a feed or search experience. For example, today on Every.org people can browse “Black-led nonprofits near Seattle” or “small nonprofits focused on climate change”. Dynamic lists in the form of the Every.org search UX can help donors with their choice overload and narrow down their selection.
The quality of the recommendations can only be as good as the data that is available. Currently, Every.org imports nonprofit information from public data and has a volunteer-driven cause tagging system that powers our search engine and feeds. It is also important that the data is as inclusive as possible and the algorithm flexible, allowing for individual preferences and choice in the resulting recommendations.
Qualitatively, there seems to be interest and demand in such dynamic “lists”, though Every.org does not use that term on our site to avoid implying that there is only a finite number of nonprofits per list. We provide a charity API used by platforms that customize usage based on their own needs.
For example, CryptoZakat uses our API to show trending Islamic nonprofits based on the number of donations. A few partners including a donor advised fund have expressed interest in using dynamically updating cause pages to show currently trending nonprofits to donors, suggesting additional demand for the technology.
Every.org will continue building out our infrastructure for dynamic curation and providing an API so that other giving experiences can benefit from a scalable method of solving for choice overload.
Conclusion
One of the major problems for curation is the time and effort required to find experts to create lists. We explored individual and organization directed lists, however most preferred to use lists for temporary fundraising. Lists will continue to serve as an effective tool for fundraising purposes, whether it be for a birthday or in response to a crisis. Our API still serves all of the lists created as part of this project, and others are free to fork our open source list client.
In the case of helping donors with choice overload, dynamically recommending nonprofits based on chosen characteristics and what is currently trending provides a scalable path forward, and is in use on our own website at www.every.org/causes and also available as the Browse endpoint in our Search API for other platforms to use.
There is more to explore in regards to curation or list-based fundraising, and we have this early prototype for creating temporary groups of nonprofits to fundraise together. We hope that both this report covering our own experiments and the open infrastructure on Every.org will provide a helpful foundation for further innovations in this space to inspire more giving.
About Every.org
Every.org is a 501(c)(3) nonprofit building an accessible giving infrastructure for a more generous world. Our goal is to empower every individual and organization to be able to use technology for good. We want to help nonprofits connect with givers and keep up with the digitization of giving — whether it’s Apple/Google Pay or cryptocurrency. Please visit every.org, learn more about us, or support our work with a tax-deductible donation.