The EA Giving Tuesday Donation Matching Initiative is supported by Rethink Charity. Rebecca Baron, Marisa Jurczyk, William Kiely, Catherine Low, and Avi Norowitz were on the organizer team for this project in 2019, and many others also provided valuable contributions.
Since 2016, Facebook has offered to match donations made on Giving Tuesday through Facebook, representing an unusual opportunity to get EA donations counterfactually matched. From 2017 there have been efforts to coordinate EA donors, especially those in the USA, to direct matching funds to EA-aligned organizations. In 2019, Facebook announced plans to match the first $7 million USD of donations starting at 8am EST Giving Tuesday (December 3, 2019).
We made a coordinated effort to direct matching funds to EA-aligned organizations, with a focus on donating fast. Despite setbacks, we had $563k USD (52%) of $1.1 million USD donations matched, comparable to the performance we saw in 2018. The matching funds ran out at 14 seconds, but donation processing delays caused some donations made significantly earlier to not get matched.
We conducted follow-up work with organizations, and found that they received the expected amounts. However, in one case, there was a delay of 5 months before an organization received their funds.
We conducted a follow-up survey with donors, which showed most donors read our instructions, spent less than 1 hour preparing for the match, but were only able to donate 73% of their intended donations on average. We also saw a substantial improvement in the experience for non-US donors, but they were still unable to donate large amounts because of donation amount limits.
We encountered challenges with last-minute changes in Facebook’s payment system which resulted in significant changes to donor instructions. We also encountered other challenges regarding delays in donation processing, difficulties in receiving funds, difficulties in data collection, and the matching program’s late announcement.
We improved our systems since 2018, including: collaborating with Rethink Charity, improvements in outreach, improvements in communication with donors, improvements in our use of technology, and more donation recipients available.
We consider a few questions for 2020 related to: the behavior of Facebook’s payment system, the match end time, donating a few seconds early, mitigating risks associated with receiving funds, our approach to donors outside the US, methods of communicating last minute changes, and CEA’s donor lotteries.
Facebook’s Giving Tuesday matching program
On November 1st, 2019, Facebook announced that they would be repeating their Giving Tuesday matching program in 2019:
Giving Tuesday (#GivingTuesday) is a global day of giving celebrated the Tuesday after Thanksgiving in the United States. To help raise awareness of charitable causes and amplify donors’ generosity, Facebook will match eligible donations up to the full match amount that are made on Facebook during Giving Tuesday 2019.Start a fundraiser for Giving Tuesday or donate to a nonprofit
Giving Tuesday 2019
- Facebook’s match begins at 8:00 AM Eastern Time (5:00 AM Pacific Time) on Tuesday, December 3, 2019.
- Donations are matched dollar for dollar on a first-come, first-served basis until the match is met.
- Any US-based 501(c)(3) nonprofit eligible to receive donations on Facebook can be matched.
- We [Facebook] cover all processing fees for donations made on Facebook to charitable organizations.
- Facebook will match up to a total of $100,000 [USD] per nonprofit organization.
- Each donor can have up to $20,000 [USD] in eligible donations matched on Giving Tuesday.
Facebook provided further details on their Social Good website. Most notably, they reported that they would match the first $7 million USD in donations.
This opportunity was most useful for US donors, who were eligible for up to $20,000 USD of matching, with each donation capped at $2499 USD. US donors were also potentially eligible for tax benefits. Donors from several other countries were able to participate, but most of these countries had total donation limits of €250, and were ineligible for tax benefits.
Facebook ran similar matching programs in 2016 for $900,000 USD, in 2017 for $2 million USD, and in 2018 for $7 million USD.
In 2018, the EA community donated $719,000 USD through our initiative, directing a counterfactual $469,000 USD in matching funds to EA charities, even though the matching funds were exhausted in 16 seconds. For more information, see our 2018 retrospective.
Our attempt in 2017 was less successful, with $48,000 USD matched out of $379,000 USD donations. However, Facebook did feature our efforts in their white paper on Giving Tuesday (search for “effective altruism”).
To learn more about the Facebook matching program and how 2019 compared to previous years, see our FAQ.
The opportunity for EAs
Facebook’s annual donation matching program on Giving Tuesday represents a rare opportunity to get donations to EA-aligned organizations counterfactually matched. This opportunity is different than most matching opportunities in the following respects:
The matching funds are available to nearly any US nonprofit organization registered as a 501(c)(3). This permits donations to EA-aligned organizations, and reduces the risk that donation decisions are distorted towards less effective organizations because donors give preference to charities that are matched.
Matching funds that are not directed to EA-aligned organizations otherwise go to organizations of more average effectiveness. This contrasts with most donation matching opportunities, where it’s likely that the matching partner would have made the same donation to the same organization anyway.
We see this donation-matching opportunity as one that avoids most of the pitfalls described by this GiveWell blog post on donation matching.
To learn more about the opportunity for EAs, see our FAQ.
Our strategy to direct the matching funds
In 2017, the matching funds ran out in 86 seconds. In 2018, that number declined to 16 seconds. In 2019, we were concerned about this trend continuing, and that the matching funds might run out in just a few seconds. Because of this concern, we emphasized the need to donate within as few seconds as possible, and preferably within the first second. To help donors accomplish this, we provided written instructions and instructional videos on donating fast.
By investigating and testing Facebook’s donation payment system, we had discovered multiple changes since 2018. We prepared a strategy to use these changes to our advantage, and our instructions to donors reflected this strategy. (Unfortunately, shortly before the match began, Facebook reverted all of these changes, which was a major setback for us. We discuss this further below.)
We also wanted to provide donors an easy way to donate to a large variety of EA-aligned organizations, and we wanted to be able to collect data on donations and matches to these organizations. To facilitate this, we created one Facebook fundraiser for each organization and added them to a spreadsheet.
Because only US registered 501(c)(3)s nonprofit organizations were eligible, we faced the risk that a large number of EA-aligned organizations, programs, and funds would remain out of reach for donors. This would have reduced the amount of donations eligible for matching, and could have influenced some donors to donate less effectively for the match. Thankfully, we were able to work with Rethink Charity, CEA, ACE, GiveWell, and other organizations to provide donors the opportunity to donate to the majority of these difficult-to-reach organizations, programs, and funds through Facebook on Giving Tuesday.
Between 2018 and 2019, the match amount limit per US registered nonprofit organization was lowered from $250k to $100k USD. This could have limited the amount of matching funds a nonprofit received, or it could have encouraged some EAs to donate less effectively while trying to stay under the match limit. To mitigate this concern, our partner Rethink Charity made use of their RC Forward platform to accept and regrant donations for EA Funds and SCI Foundation. This helped CEA and GiveWell remain under their $100k USD match limit. In addition, we coordinated with as many organizations as we could to help them receive donations directly rather than receiving donations through a regranting organizations that could have reached the $100k USD limit.
Even though we only received confirmation about a matching program 1 month in advance, we began early preparations for the match 3 months in advance. This gave us the opportunity to carefully plan our strategy, give organizations a heads up, promote the opportunity to donors, arrange a plan to get donations and matches to difficult-to-reach organizations, and prepare quality materials.
By the Saturday morning before Giving Tuesday, we had $500k USD in pledges. By the time the match began on Tuesday morning, this had increased to $1.3 million USD in pledges from 250 pledgers, with a median pledge of $1500 USD
After our fundraisers ended at midnight on Wednesday, we had $1.1 million USD in donations from 330 donors, with a median amount per donor of $1000 USD. $938k USD of these donations were made within the first minute after the match began.
It’s notable that donation amounts were lower than pledge amounts, even though there were more donors than pledgers. This may have been caused by the changes Facebook made to their payment system shortly before the match began. These changes made it more difficult to donate large amounts within the first few seconds after the match began.
Organizations on Facebook Payments should have received donation amounts in early January 2020. Other organizations should have received donation amounts in late February 2020, in the form of a check or an electronic funds transfer from Network For Good.
Match amounts
On January 21st, 2020, Facebook reported the match amounts. $563k USD (52%) of donations were matched
You can find a more complete breakdown of donations and matches on our website.
Organizations should have received match amounts in late February 2020 as a check or an electronic funds transfer from Network For Good.
Match end time
While we encouraged donors to click the donate button the moment the match time started, donors were restricted to $2499 USD per donation, so those wanting to donate more than this amount had to make multiple donations before the match ran out.
Organizations registered with Facebook Payments can generate a “Daily Transactions Report”, which includes a “Charge Time” column. Based on the information we collected, we estimate that the match ended during second 14, as measured by the “Charge Time”.
During our testing before the match began, we saw a 1 to 3 second delay between the click of the final donate button and the “Charge Time”. However, as soon as the match began, many donors experienced processing delays that were considerably longer. This meant that even some donations made with a click within the first few seconds may have been processed after the match had already ended at 14 seconds.
If we take the 14 second match end time at face value, which is only two seconds earlier than the 16 seconds we saw in 2018 (as measured by the “Charge Time”), this could lead to the interpretation that the match did not become substantially more competitive between 2018 and 2019. However, because many donations made with a click before 14 seconds were not matched, this interpretation would be misleading.
Unfortunately, it’s not possible to determine when the 2019 match ended in terms of “click time,” because (a) we don’t have reliable data on when people clicked donate, and (b) the time would vary by donation. However, because processing delays in 2018 were minimal, we do have reliable data on when donors clicked the donate buttons 2018. If we assume that donors in 2019 donated as fast as donors in 2018, then this might imply a match end time of 7 seconds as measured by average “click time.”
Follow-up with organizations
We conducted a follow-up with organizations that we expected to receive $1000 USD or more in donations and/or matches. We provided them our estimated total donation and match amounts, as well as the timestamps, donor names, donation amounts, and match amounts for each donation. Since organizations generally aren’t able to see contact information for Facebook donors, we’re also providing contact information for pledgers who gave us permission to do so.
We’ve also asked organizations to provide us with the donation and match amounts they received on Facebook on Giving Tuesday, and we’ve updated our website with the amounts reported. Overall, the donation and match amounts reported by organizations have been consistent with our expectations.
We’ve also followed up with organizations regarding donations that need to be restricted or regranted to ensure that the amounts are allocated correctly.
Follow-up with donors
At around 20 minutes after the match began, we sent a follow-up survey to donors. We performed an analysis on the survey results, which you can find here. We seem to have captured most donors in our survey, weighted donation amounts or number of donors. The following are some of the key takeaways from the survey:
Most small donors spent less than 1 hour on preparation, while most medium and large donors spent less than 2 hours.
Nearly all donors reported reading our instructions.
Donors were only able to donate 73% of their intended amount, most likely because of unexpected changes in Facebook’s payment limits and other payment problems.
Although 26% of donors were outside the US, they only accounted for 5% of donated amounts, most likely because of low donation amount limits outside the US.
A minority of donors report donating to a different organization because of the matching opportunity, but very few reported donating to a less effective organization.
13% of donors reported losing the ability to donate appreciated shares, weighted by donation amount. There were few other reports of lost tax benefits because of the matching opportunity.
Weighted by donation amounts, we found that 37% of donors “definitely” or “probably” would have donated on Facebook with the intention to get matched without our work.
In 2018, donors reported being motivated to donate 12% more in the short term because of the matching opportunity. If we extrapolate this to donations in 2019, this amounts to $131,000 USD.
Our observations of payment problems were comparable to 2018, with one notable improvement being a substantial reduction in payment problems outside of the US.
For more details, see our full survey analysis here.
In addition to our follow-up survey, when Facebook began reporting match amounts, we sent a follow-up e-mail to donors to inform them of (a) the total amounts matched and (b) instructions to check whether their own donations were matched.
Challenges
In this section, we review some of the unexpected challenges we faced.
Unexpected changes in Facebook’s payment system
Early on in our planning, we observed the following behavior for US donors in Facebook’s payment system:
We observed a soft limit on donations of $2,499 USD. We found that this soft limit could be lifted in advance by triggering and filling out a “Confirm your identity” dialog.
We observed that for donations $500 USD and greater, the donor is prompted with an additional “Confirm your donation” dialog before the donations is processed.
With a goal of helping EAs donate as fast as possible, we included the following strategies in our instructions to US donors:
US donors intending to donate more than $2,499 USD were provided instructions to trigger and complete the “Confirm your identity” dialog in advance, permitting up to $9,999 USD per donation.
US donors intending to donate $500 USD or more were provided instructions to click the “Donate” button before the match to trigger the “Confirm your donation” dialog early, so that the final donation could be made with a single click.
We had some concern that Facebook might change the behaviors described above, or other behaviors. To mitigate this risk, we had team members and volunteers frequently running tests of Facebook’s payment system, increasing the frequency of tests as the match start time approached.
Unfortunately, the “Confirm your donation” dialog stopped appearing for US donors around a day before the match began, which required us to quickly change our instructions and communicate those changes to donors. To further complicate matters, we didn’t know whether this change also affected international donors, which required further testing and additional work in a short amount of time.
To make matters worse, around 30 minutes before the match began, the US donation amount hard limit was reduced to $2,499 USD, making the “Confirm your identity” dialog instructions irrelevant. In effect, Facebook’s payment system reverted to the behavior we saw in 2018, which required 8 separate donations to approach $20k USD in donations. With the short time frame we had, we communicated this change to donors as fast as we could via e-mail, our website, our fundraisers spreadsheet, and Facebook.
Despite these challenges, EAs were still able to donate fast enough to get $563k USD (52%) of their donations matched.
If we had known what Facebook was intending, our instructions would have asked large donors to practice making multiple donations in quick succession, so it is likely that the match amount would have been significantly higher. In addition, these last minute changes may have caused larger donors to donate less than they initially intended.
Delays in donation processing
As discussed earlier, many donors experienced processing delays of several seconds. Although this was frustrating for many donors, the effect on the overall amounts matched may not be as large as it may initially seem. Other fast donors experienced the same processing delays as EAs, and without the processing delays, the match would have ended much earlier.
Difficulty in receiving funds
Organizations not signed up for Facebook Payments receive both donated and match amounts through Network For Good, a donor advised fund. We’ve been following up with organizations each year to ensure they’re receiving the funds expected, and we had previously encountered no significant indicators of problems in this area.
Unfortunately, the Centre for Effective Altruism this year had their $229k USD in funds flagged for review by Network For Good. This led to an additional delay of 2 months in receiving the funds, with little communication from Network For Good during that period. Fortunately, Network For Good released the funds in early May 2020, and the Centre for Effective Altruism will be restricting and regranting the funds as appropriate.
Difficulties in data collection
We collect data on donations and matches largely from Facebook fundraiser pages. We had an incident in which Facebook had taken our fundraiser pages for the Centre for Effective Altruism offline due to erroneous spam detection. Although we had collected the data from the fundraiser page, we could have missed a few later donations. In addition, we could have faced difficulty obtaining data on the match amounts for each of those fundraisers. Although we had methods of reconstructing most of the missing data, this could have led to reduced accuracy in how donations and matches were allocated.
Fortunately, we were able to work with Facebook to get the fundraiser pages restored, and we’ve been able to collect all the data we needed.
The erroneous spam detection was triggered by our posting of the donor follow-up survey to a large number of our fundraisers, which used the same URL shortener domain (eagiv.org) that we had used in the descriptions of the affected fundraiser pages. Moving forward, we will likely avoid posting the follow-up survey on fundraiser pages.
Late announcement of the matching program
A challenge we face each year is that we don’t know if and when Facebook will announce a matching program or its details. For Giving Tuesday 2019, Facebook announced their matching program only 32 days in advance, giving us far less time than the 48 days given in the previous year.
Our general approach has been to start planning and outreach for the matching program a few months in advance, working on the assumption that the details will be roughly similar to the details of the previous year. Although we risk wasting time and building publicity for a matching program that doesn’t materialize, we believe the benefits of advance planning make this worth it.
Areas of improvement since 2018
Each year we try to learn from our experience in previous years and increase our performance. This section reviews some of the areas we’ve improved in 2019 compared to 2018.
Collaboration with Rethink Charity
After our success in 2018, we were looking for ways to improve our effectiveness. We approached Rethink Charity in September 2019 to ask if they’d be interested in collaborating with us to increase the scale, sustainability, and professionalism of the project. Rethink Charity was happy to collaborate, and supported the project in a variety of ways. This included:
Rethink Charity assigned us two of their team members, Catherine Low and Marisa Jurczyk, to contribute a large number of hours towards the project.
Rethink Charity, making use of their RC Forward infrastructure, led by Siobhan Brenton, offered to accept and regrant donations for multiple EA Funds and SCI to maximize the amount of possible matching funds towards EA causes.
We look forward to including our collaboration with Rethink Charity in future years.
Improved outreach and marketing
This year, we substantially expanded our efforts to promote the opportunity to donors in the EA community. We began promotional work in early October 2019 and continued until Giving Tuesday itself. We made use of sign-up forms, e-mail announcements, and Facebook. We tried to reach donors directly, as well as group organizers who could share the opportunity with their members.
In addition, we also expanded our more individualized outreach to large and medium donors, requesting advice from professional fundraisers, and reaching out 164 donors individually.
Improved communications with donors
For Giving Tuesday 2018, we made the mistake of not emailing donors to inform them most donations had been matched. Since Facebook did not communicate this either, we heard feedback from some donors that they had assumed their donations were not matched.
For Giving Tuesday 2019, we communicated to donors on both Facebook and email that $563k USD (52%) of EA donations had been matched, and provided instructions for them to check if their own donations had been matched.
Improved use of technology
This year we made use of a number of additional technology tools to address specific limitations we were facing:
HubSpot, for a shared email inbox and a ticketing system.
MailChimp, for email announcements.
G-Suite, for email.
In addition, we applied a workaround for a major problem we were having with social media previews of our website.
More organizations, funds, and programs available
This year we were able to make more organizations, funds, and programs available for donation, increasing the count from 64 to 81. In some of the more challenging cases, we were able to permit donations by coordinating with one organization to accept restricted donations to be regranted to others.
Questions for 2020
If the Facebook donation matching opportunity is available again in 2020, our experience in 2019 raises some questions for 2020.
What behavior should we expect on Facebook’s payment system?
As described above, one of the challenges we face is that Facebook could change the behavior of their payment system at any time before the match begins. This makes it difficult for us to provide reliable instructions to donors on placing their donations fast. While it seems likely the behavior we saw during the 2019 match will be in place during the 2020 match, we’ll need a plan that accounts for the possibility of different behavior.
How much time will we have?
In 2016, Facebook reported that the match ran out “within hours,” although we don’t have independent evidence of this. In 2017 the match ended at 86 seconds, and in 2018 the match ended at 16 seconds. In 2019 the match ended at 14 seconds as measured by donation processing time, but it seems likely that because of donation processing delays, most donations made after 7 seconds were not matched.
If trends continue in 2020, the match could run out even faster. Our strategy will likely vary depending on how fast we expect the matching funds to run out.
Should we recommend donating early?
In our testing before the 2019 match, we observed 2 to 3 second delays in donation processing times, as measured by the “Charge Time” on the “Daily Transaction Report” that organizations on Facebook Payments can generate. We believed that “Charge Time” was used for matching eligibility, and the outcome of tests we conducted during the 2019 match provided further evidence for that. Therefore, we may be able to increase our match percent in 2020 by asking donors to start donating 1 to 2 seconds before the match start time. This may be particularly beneficial for large donors, who will likely be attempting 8 individual donations before the match ends.
However, there’s some risk that donations made 1 to 2 seconds early would not be eligible for matching. For instance, although Facebook seems to have used the “Charge Time” for determining matching eligibility in 2018 and 2019, it’s possible Facebook could use a different timestamp in 2020 with a smaller delay. (We are aware of at least 3 kinds of timestamps, and there may be others we are not aware of.)
How can we mitigate risks associated with receiving funds?
As noted above, the Centre for Effective Altruism experienced delays in receiving their Giving Tuesday 2019 funds from Network For Good. Although Network For Good released the funds in May 2019, the experience provides reason for concern. Therefore, we should consider how we could mitigate associated risks moving forward.
Although organizations can only receive matching funds through Network For Good, they have the option to receive donations more directly through Facebook Payments. So we may want to advise organizations to sign up for Facebook Payments in advance. We’ve discovered, however, that signing up for Facebook Payments too close to Giving Tuesday could result in organizations being ineligible to receive donations on Giving Tuesday, so we’d need to ensure that these sign-ups are done well in advance. In addition, it’s possible that receiving donations through Facebook Payments could complicate administrative work for some organizations.
How much should we focus on donors outside the US?
In the 2018 survey, donors outside the US reported a very large number of payment problems. We had a goal in 2019 to improve the experience of donors outside the US and increase overall matching funds. We put a considerable amount of effort into recruiting volunteers outside of the US to help test Facebook’s payment limits, and prepared instructions for donors in different countries. Our 2019 survey results show that we were highly successful at reducing payment problems among donors outside the US. Unfortunately, although donors outside the US represented 26% of donors, they still accounted for only 5% of donation amounts, most likely because of the very low donation amount limits that most non-US donors faced. This raises the question of whether we should reduce our focus on supporting donors outside the US relative to other areas.
How can we improve communications of last-minute changes?
As noted above, Facebook changed the behavior of their payment system 30 minutes before the match began. This required that donations of amounts larger than $2,499 USD needed to be split into multiple smaller donations. We communicated this change as effectively as we could given the circumstances, and the amounts matched suggest we did reasonably well. However, we still may want to consider if there are other ways we could communicate last minute changes more effectively.
Can we include CEA’s donor lotteries?
The day before Giving Tuesday 2019, a few donors expressed interest in donating to CEA’s donor lotteries through Facebook, with the opportunity to get those donations matched. CEA also expressed some interest in this. Unfortunately, given the time constraints and complexities, it was not possible to make this work.
Moving forward, it’s still not clear whether it will be feasible to integrate CEA’s donor lottery into Facebook’s Giving Tuesday matching program in 2020. It introduces a number of complexities and risks, and we’d need to consider the additional administrative overhead that may be required for our team and for CEA. However, we’ll likely revisit this as we approach Giving Tuesday 2020. We would like to include as many EA donation opportunities as feasible.
EA Giving Tuesday Donation Matching Initiative 2019 Retrospective
The EA Giving Tuesday Donation Matching Initiative is supported by Rethink Charity. Rebecca Baron, Marisa Jurczyk, William Kiely, Catherine Low, and Avi Norowitz were on the organizer team for this project in 2019, and many others also provided valuable contributions.
To stay up to date about EA Giving Tuesday 2020, please sign-up on our website at https://www.eagivingtuesday.org.
Summary
Since 2016, Facebook has offered to match donations made on Giving Tuesday through Facebook, representing an unusual opportunity to get EA donations counterfactually matched. From 2017 there have been efforts to coordinate EA donors, especially those in the USA, to direct matching funds to EA-aligned organizations. In 2019, Facebook announced plans to match the first $7 million USD of donations starting at 8am EST Giving Tuesday (December 3, 2019).
We made a coordinated effort to direct matching funds to EA-aligned organizations, with a focus on donating fast. Despite setbacks, we had $563k USD (52%) of $1.1 million USD donations matched, comparable to the performance we saw in 2018. The matching funds ran out at 14 seconds, but donation processing delays caused some donations made significantly earlier to not get matched.
We conducted follow-up work with organizations, and found that they received the expected amounts. However, in one case, there was a delay of 5 months before an organization received their funds.
We conducted a follow-up survey with donors, which showed most donors read our instructions, spent less than 1 hour preparing for the match, but were only able to donate 73% of their intended donations on average. We also saw a substantial improvement in the experience for non-US donors, but they were still unable to donate large amounts because of donation amount limits.
We encountered challenges with last-minute changes in Facebook’s payment system which resulted in significant changes to donor instructions. We also encountered other challenges regarding delays in donation processing, difficulties in receiving funds, difficulties in data collection, and the matching program’s late announcement.
We improved our systems since 2018, including: collaborating with Rethink Charity, improvements in outreach, improvements in communication with donors, improvements in our use of technology, and more donation recipients available.
We consider a few questions for 2020 related to: the behavior of Facebook’s payment system, the match end time, donating a few seconds early, mitigating risks associated with receiving funds, our approach to donors outside the US, methods of communicating last minute changes, and CEA’s donor lotteries.
Facebook’s Giving Tuesday matching program
On November 1st, 2019, Facebook announced that they would be repeating their Giving Tuesday matching program in 2019:
Facebook provided further details on their Social Good website. Most notably, they reported that they would match the first $7 million USD in donations.
This opportunity was most useful for US donors, who were eligible for up to $20,000 USD of matching, with each donation capped at $2499 USD. US donors were also potentially eligible for tax benefits. Donors from several other countries were able to participate, but most of these countries had total donation limits of €250, and were ineligible for tax benefits.
Facebook ran similar matching programs in 2016 for $900,000 USD, in 2017 for $2 million USD, and in 2018 for $7 million USD.
In 2018, the EA community donated $719,000 USD through our initiative, directing a counterfactual $469,000 USD in matching funds to EA charities, even though the matching funds were exhausted in 16 seconds. For more information, see our 2018 retrospective.
Our attempt in 2017 was less successful, with $48,000 USD matched out of $379,000 USD donations. However, Facebook did feature our efforts in their white paper on Giving Tuesday (search for “effective altruism”).
To learn more about the Facebook matching program and how 2019 compared to previous years, see our FAQ.
The opportunity for EAs
Facebook’s annual donation matching program on Giving Tuesday represents a rare opportunity to get donations to EA-aligned organizations counterfactually matched. This opportunity is different than most matching opportunities in the following respects:
The matching funds are available to nearly any US nonprofit organization registered as a 501(c)(3). This permits donations to EA-aligned organizations, and reduces the risk that donation decisions are distorted towards less effective organizations because donors give preference to charities that are matched.
Matching funds that are not directed to EA-aligned organizations otherwise go to organizations of more average effectiveness. This contrasts with most donation matching opportunities, where it’s likely that the matching partner would have made the same donation to the same organization anyway.
We see this donation-matching opportunity as one that avoids most of the pitfalls described by this GiveWell blog post on donation matching.
To learn more about the opportunity for EAs, see our FAQ.
Our strategy to direct the matching funds
In 2017, the matching funds ran out in 86 seconds. In 2018, that number declined to 16 seconds. In 2019, we were concerned about this trend continuing, and that the matching funds might run out in just a few seconds. Because of this concern, we emphasized the need to donate within as few seconds as possible, and preferably within the first second. To help donors accomplish this, we provided written instructions and instructional videos on donating fast.
By investigating and testing Facebook’s donation payment system, we had discovered multiple changes since 2018. We prepared a strategy to use these changes to our advantage, and our instructions to donors reflected this strategy. (Unfortunately, shortly before the match began, Facebook reverted all of these changes, which was a major setback for us. We discuss this further below.)
We also wanted to provide donors an easy way to donate to a large variety of EA-aligned organizations, and we wanted to be able to collect data on donations and matches to these organizations. To facilitate this, we created one Facebook fundraiser for each organization and added them to a spreadsheet.
Because only US registered 501(c)(3)s nonprofit organizations were eligible, we faced the risk that a large number of EA-aligned organizations, programs, and funds would remain out of reach for donors. This would have reduced the amount of donations eligible for matching, and could have influenced some donors to donate less effectively for the match. Thankfully, we were able to work with Rethink Charity, CEA, ACE, GiveWell, and other organizations to provide donors the opportunity to donate to the majority of these difficult-to-reach organizations, programs, and funds through Facebook on Giving Tuesday.
Between 2018 and 2019, the match amount limit per US registered nonprofit organization was lowered from $250k to $100k USD. This could have limited the amount of matching funds a nonprofit received, or it could have encouraged some EAs to donate less effectively while trying to stay under the match limit. To mitigate this concern, our partner Rethink Charity made use of their RC Forward platform to accept and regrant donations for EA Funds and SCI Foundation. This helped CEA and GiveWell remain under their $100k USD match limit. In addition, we coordinated with as many organizations as we could to help them receive donations directly rather than receiving donations through a regranting organizations that could have reached the $100k USD limit.
Even though we only received confirmation about a matching program 1 month in advance, we began early preparations for the match 3 months in advance. This gave us the opportunity to carefully plan our strategy, give organizations a heads up, promote the opportunity to donors, arrange a plan to get donations and matches to difficult-to-reach organizations, and prepare quality materials.
To learn more about our strategy, see our FAQ.
Pledge and donation amounts
By the Saturday morning before Giving Tuesday, we had $500k USD in pledges. By the time the match began on Tuesday morning, this had increased to $1.3 million USD in pledges from 250 pledgers, with a median pledge of $1500 USD
After our fundraisers ended at midnight on Wednesday, we had $1.1 million USD in donations from 330 donors, with a median amount per donor of $1000 USD. $938k USD of these donations were made within the first minute after the match began.
It’s notable that donation amounts were lower than pledge amounts, even though there were more donors than pledgers. This may have been caused by the changes Facebook made to their payment system shortly before the match began. These changes made it more difficult to donate large amounts within the first few seconds after the match began.
You can see a more complete breakdown of pledges and donations on our website.
Organizations on Facebook Payments should have received donation amounts in early January 2020. Other organizations should have received donation amounts in late February 2020, in the form of a check or an electronic funds transfer from Network For Good.
Match amounts
On January 21st, 2020, Facebook reported the match amounts. $563k USD (52%) of donations were matched
You can find a more complete breakdown of donations and matches on our website.
Organizations should have received match amounts in late February 2020 as a check or an electronic funds transfer from Network For Good.
Match end time
While we encouraged donors to click the donate button the moment the match time started, donors were restricted to $2499 USD per donation, so those wanting to donate more than this amount had to make multiple donations before the match ran out.
Organizations registered with Facebook Payments can generate a “Daily Transactions Report”, which includes a “Charge Time” column. Based on the information we collected, we estimate that the match ended during second 14, as measured by the “Charge Time”.
During our testing before the match began, we saw a 1 to 3 second delay between the click of the final donate button and the “Charge Time”. However, as soon as the match began, many donors experienced processing delays that were considerably longer. This meant that even some donations made with a click within the first few seconds may have been processed after the match had already ended at 14 seconds.
If we take the 14 second match end time at face value, which is only two seconds earlier than the 16 seconds we saw in 2018 (as measured by the “Charge Time”), this could lead to the interpretation that the match did not become substantially more competitive between 2018 and 2019. However, because many donations made with a click before 14 seconds were not matched, this interpretation would be misleading.
Unfortunately, it’s not possible to determine when the 2019 match ended in terms of “click time,” because (a) we don’t have reliable data on when people clicked donate, and (b) the time would vary by donation. However, because processing delays in 2018 were minimal, we do have reliable data on when donors clicked the donate buttons 2018. If we assume that donors in 2019 donated as fast as donors in 2018, then this might imply a match end time of 7 seconds as measured by average “click time.”
Follow-up with organizations
We conducted a follow-up with organizations that we expected to receive $1000 USD or more in donations and/or matches. We provided them our estimated total donation and match amounts, as well as the timestamps, donor names, donation amounts, and match amounts for each donation. Since organizations generally aren’t able to see contact information for Facebook donors, we’re also providing contact information for pledgers who gave us permission to do so.
We’ve also asked organizations to provide us with the donation and match amounts they received on Facebook on Giving Tuesday, and we’ve updated our website with the amounts reported. Overall, the donation and match amounts reported by organizations have been consistent with our expectations.
We’ve also followed up with organizations regarding donations that need to be restricted or regranted to ensure that the amounts are allocated correctly.
Follow-up with donors
At around 20 minutes after the match began, we sent a follow-up survey to donors. We performed an analysis on the survey results, which you can find here. We seem to have captured most donors in our survey, weighted donation amounts or number of donors. The following are some of the key takeaways from the survey:
Most small donors spent less than 1 hour on preparation, while most medium and large donors spent less than 2 hours.
Nearly all donors reported reading our instructions.
Donors were only able to donate 73% of their intended amount, most likely because of unexpected changes in Facebook’s payment limits and other payment problems.
Although 26% of donors were outside the US, they only accounted for 5% of donated amounts, most likely because of low donation amount limits outside the US.
A minority of donors report donating to a different organization because of the matching opportunity, but very few reported donating to a less effective organization.
13% of donors reported losing the ability to donate appreciated shares, weighted by donation amount. There were few other reports of lost tax benefits because of the matching opportunity.
Weighted by donation amounts, we found that 37% of donors “definitely” or “probably” would have donated on Facebook with the intention to get matched without our work.
In 2018, donors reported being motivated to donate 12% more in the short term because of the matching opportunity. If we extrapolate this to donations in 2019, this amounts to $131,000 USD.
Our observations of payment problems were comparable to 2018, with one notable improvement being a substantial reduction in payment problems outside of the US.
For more details, see our full survey analysis here.
In addition to our follow-up survey, when Facebook began reporting match amounts, we sent a follow-up e-mail to donors to inform them of (a) the total amounts matched and (b) instructions to check whether their own donations were matched.
Challenges
In this section, we review some of the unexpected challenges we faced.
Unexpected changes in Facebook’s payment system
Early on in our planning, we observed the following behavior for US donors in Facebook’s payment system:
We observed a soft limit on donations of $2,499 USD. We found that this soft limit could be lifted in advance by triggering and filling out a “Confirm your identity” dialog.
We observed that for donations $500 USD and greater, the donor is prompted with an additional “Confirm your donation” dialog before the donations is processed.
With a goal of helping EAs donate as fast as possible, we included the following strategies in our instructions to US donors:
US donors intending to donate more than $2,499 USD were provided instructions to trigger and complete the “Confirm your identity” dialog in advance, permitting up to $9,999 USD per donation.
US donors intending to donate $500 USD or more were provided instructions to click the “Donate” button before the match to trigger the “Confirm your donation” dialog early, so that the final donation could be made with a single click.
We had some concern that Facebook might change the behaviors described above, or other behaviors. To mitigate this risk, we had team members and volunteers frequently running tests of Facebook’s payment system, increasing the frequency of tests as the match start time approached.
Unfortunately, the “Confirm your donation” dialog stopped appearing for US donors around a day before the match began, which required us to quickly change our instructions and communicate those changes to donors. To further complicate matters, we didn’t know whether this change also affected international donors, which required further testing and additional work in a short amount of time.
To make matters worse, around 30 minutes before the match began, the US donation amount hard limit was reduced to $2,499 USD, making the “Confirm your identity” dialog instructions irrelevant. In effect, Facebook’s payment system reverted to the behavior we saw in 2018, which required 8 separate donations to approach $20k USD in donations. With the short time frame we had, we communicated this change to donors as fast as we could via e-mail, our website, our fundraisers spreadsheet, and Facebook.
Despite these challenges, EAs were still able to donate fast enough to get $563k USD (52%) of their donations matched.
If we had known what Facebook was intending, our instructions would have asked large donors to practice making multiple donations in quick succession, so it is likely that the match amount would have been significantly higher. In addition, these last minute changes may have caused larger donors to donate less than they initially intended.
Delays in donation processing
As discussed earlier, many donors experienced processing delays of several seconds. Although this was frustrating for many donors, the effect on the overall amounts matched may not be as large as it may initially seem. Other fast donors experienced the same processing delays as EAs, and without the processing delays, the match would have ended much earlier.
Difficulty in receiving funds
Organizations not signed up for Facebook Payments receive both donated and match amounts through Network For Good, a donor advised fund. We’ve been following up with organizations each year to ensure they’re receiving the funds expected, and we had previously encountered no significant indicators of problems in this area.
Unfortunately, the Centre for Effective Altruism this year had their $229k USD in funds flagged for review by Network For Good. This led to an additional delay of 2 months in receiving the funds, with little communication from Network For Good during that period. Fortunately, Network For Good released the funds in early May 2020, and the Centre for Effective Altruism will be restricting and regranting the funds as appropriate.
Difficulties in data collection
We collect data on donations and matches largely from Facebook fundraiser pages. We had an incident in which Facebook had taken our fundraiser pages for the Centre for Effective Altruism offline due to erroneous spam detection. Although we had collected the data from the fundraiser page, we could have missed a few later donations. In addition, we could have faced difficulty obtaining data on the match amounts for each of those fundraisers. Although we had methods of reconstructing most of the missing data, this could have led to reduced accuracy in how donations and matches were allocated.
Fortunately, we were able to work with Facebook to get the fundraiser pages restored, and we’ve been able to collect all the data we needed.
The erroneous spam detection was triggered by our posting of the donor follow-up survey to a large number of our fundraisers, which used the same URL shortener domain (eagiv.org) that we had used in the descriptions of the affected fundraiser pages. Moving forward, we will likely avoid posting the follow-up survey on fundraiser pages.
Late announcement of the matching program
A challenge we face each year is that we don’t know if and when Facebook will announce a matching program or its details. For Giving Tuesday 2019, Facebook announced their matching program only 32 days in advance, giving us far less time than the 48 days given in the previous year.
Our general approach has been to start planning and outreach for the matching program a few months in advance, working on the assumption that the details will be roughly similar to the details of the previous year. Although we risk wasting time and building publicity for a matching program that doesn’t materialize, we believe the benefits of advance planning make this worth it.
Areas of improvement since 2018
Each year we try to learn from our experience in previous years and increase our performance. This section reviews some of the areas we’ve improved in 2019 compared to 2018.
Collaboration with Rethink Charity
After our success in 2018, we were looking for ways to improve our effectiveness. We approached Rethink Charity in September 2019 to ask if they’d be interested in collaborating with us to increase the scale, sustainability, and professionalism of the project. Rethink Charity was happy to collaborate, and supported the project in a variety of ways. This included:
Rethink Charity assigned us two of their team members, Catherine Low and Marisa Jurczyk, to contribute a large number of hours towards the project.
Rethink Charity, making use of their RC Forward infrastructure, led by Siobhan Brenton, offered to accept and regrant donations for multiple EA Funds and SCI to maximize the amount of possible matching funds towards EA causes.
We look forward to including our collaboration with Rethink Charity in future years.
Improved outreach and marketing
This year, we substantially expanded our efforts to promote the opportunity to donors in the EA community. We began promotional work in early October 2019 and continued until Giving Tuesday itself. We made use of sign-up forms, e-mail announcements, and Facebook. We tried to reach donors directly, as well as group organizers who could share the opportunity with their members.
In addition, we also expanded our more individualized outreach to large and medium donors, requesting advice from professional fundraisers, and reaching out 164 donors individually.
Improved communications with donors
For Giving Tuesday 2018, we made the mistake of not emailing donors to inform them most donations had been matched. Since Facebook did not communicate this either, we heard feedback from some donors that they had assumed their donations were not matched.
For Giving Tuesday 2019, we communicated to donors on both Facebook and email that $563k USD (52%) of EA donations had been matched, and provided instructions for them to check if their own donations had been matched.
Improved use of technology
This year we made use of a number of additional technology tools to address specific limitations we were facing:
HubSpot, for a shared email inbox and a ticketing system.
MailChimp, for email announcements.
G-Suite, for email.
In addition, we applied a workaround for a major problem we were having with social media previews of our website.
More organizations, funds, and programs available
This year we were able to make more organizations, funds, and programs available for donation, increasing the count from 64 to 81. In some of the more challenging cases, we were able to permit donations by coordinating with one organization to accept restricted donations to be regranted to others.
Questions for 2020
If the Facebook donation matching opportunity is available again in 2020, our experience in 2019 raises some questions for 2020.
What behavior should we expect on Facebook’s payment system?
As described above, one of the challenges we face is that Facebook could change the behavior of their payment system at any time before the match begins. This makes it difficult for us to provide reliable instructions to donors on placing their donations fast. While it seems likely the behavior we saw during the 2019 match will be in place during the 2020 match, we’ll need a plan that accounts for the possibility of different behavior.
How much time will we have?
In 2016, Facebook reported that the match ran out “within hours,” although we don’t have independent evidence of this. In 2017 the match ended at 86 seconds, and in 2018 the match ended at 16 seconds. In 2019 the match ended at 14 seconds as measured by donation processing time, but it seems likely that because of donation processing delays, most donations made after 7 seconds were not matched.
If trends continue in 2020, the match could run out even faster. Our strategy will likely vary depending on how fast we expect the matching funds to run out.
Should we recommend donating early?
In our testing before the 2019 match, we observed 2 to 3 second delays in donation processing times, as measured by the “Charge Time” on the “Daily Transaction Report” that organizations on Facebook Payments can generate. We believed that “Charge Time” was used for matching eligibility, and the outcome of tests we conducted during the 2019 match provided further evidence for that. Therefore, we may be able to increase our match percent in 2020 by asking donors to start donating 1 to 2 seconds before the match start time. This may be particularly beneficial for large donors, who will likely be attempting 8 individual donations before the match ends.
However, there’s some risk that donations made 1 to 2 seconds early would not be eligible for matching. For instance, although Facebook seems to have used the “Charge Time” for determining matching eligibility in 2018 and 2019, it’s possible Facebook could use a different timestamp in 2020 with a smaller delay. (We are aware of at least 3 kinds of timestamps, and there may be others we are not aware of.)
How can we mitigate risks associated with receiving funds?
As noted above, the Centre for Effective Altruism experienced delays in receiving their Giving Tuesday 2019 funds from Network For Good. Although Network For Good released the funds in May 2019, the experience provides reason for concern. Therefore, we should consider how we could mitigate associated risks moving forward.
Although organizations can only receive matching funds through Network For Good, they have the option to receive donations more directly through Facebook Payments. So we may want to advise organizations to sign up for Facebook Payments in advance. We’ve discovered, however, that signing up for Facebook Payments too close to Giving Tuesday could result in organizations being ineligible to receive donations on Giving Tuesday, so we’d need to ensure that these sign-ups are done well in advance. In addition, it’s possible that receiving donations through Facebook Payments could complicate administrative work for some organizations.
How much should we focus on donors outside the US?
In the 2018 survey, donors outside the US reported a very large number of payment problems. We had a goal in 2019 to improve the experience of donors outside the US and increase overall matching funds. We put a considerable amount of effort into recruiting volunteers outside of the US to help test Facebook’s payment limits, and prepared instructions for donors in different countries. Our 2019 survey results show that we were highly successful at reducing payment problems among donors outside the US. Unfortunately, although donors outside the US represented 26% of donors, they still accounted for only 5% of donation amounts, most likely because of the very low donation amount limits that most non-US donors faced. This raises the question of whether we should reduce our focus on supporting donors outside the US relative to other areas.
How can we improve communications of last-minute changes?
As noted above, Facebook changed the behavior of their payment system 30 minutes before the match began. This required that donations of amounts larger than $2,499 USD needed to be split into multiple smaller donations. We communicated this change as effectively as we could given the circumstances, and the amounts matched suggest we did reasonably well. However, we still may want to consider if there are other ways we could communicate last minute changes more effectively.
Can we include CEA’s donor lotteries?
The day before Giving Tuesday 2019, a few donors expressed interest in donating to CEA’s donor lotteries through Facebook, with the opportunity to get those donations matched. CEA also expressed some interest in this. Unfortunately, given the time constraints and complexities, it was not possible to make this work.
Moving forward, it’s still not clear whether it will be feasible to integrate CEA’s donor lottery into Facebook’s Giving Tuesday matching program in 2020. It introduces a number of complexities and risks, and we’d need to consider the additional administrative overhead that may be required for our team and for CEA. However, we’ll likely revisit this as we approach Giving Tuesday 2020. We would like to include as many EA donation opportunities as feasible.