This post provides summary metrics we look at to gauge GiveWell’s influence and growth. Our full annual review and plan will follow over the next few days.
- We tracked over $1.5 million in donations to top charities in 2010, compared to just over $1 million in 2009.
- Our website traffic nearly doubled from 2009 to 2010, and donations through the website nearly tripled. Our overall increase in money moved appears to be driven mostly by (a) a gain of $200,000 in six-figure donations; (b) new donors, largely acquired via search engine traffic and the outreach of Peter Singer.
- Our growth in online donations to recommended charities was significantly faster than that of the more established online donor resources (Charity Navigator and GuideStar); our total online donations remain lower than these resources’, but are now in the same ballpark.
Total money moved
One of our primary metrics is “money moved”: donations to our top charities that we can confidently identify as being made on the strength of our recommendation. We are generally very conservative in identifying “money moved”; full details of how we track it are at the end of this post.
The tables below show GiveWell’s 2010 money moved by (a) organization and (b) size of donation. They show a total of just over $1.5 million, the vast majority of which went to our top-rated charity, VillageReach.
Note: these figures do not match up exactly to the above total money moved figures because this table does not include donations made directly to our recommended charities that the charities informed us about and for which we do not know the individual donation sizes.
Donations through the website
While the aggregate money moved figure (which includes all money donated to charities due to GiveWell’s recommendation) is ultimately the more meaningful measure of GiveWell’s impact, we believe that donations by website visitors whom GiveWell staff don’t know personally is a more meaningful measure of GiveWell’s progress as these donations represent (a) use of the GiveWell tool as we ultimately envision it (i.e., retail donors coming to the site and using it to give) and (b) are less susceptible to large, one-off circumstances that lead to large variations in the aggregate money moved figure.
The following charts show monthly donations through the website, for 2008-2010. (We include January 2011 data.) We’ve attempted to strip out donors we have had personal contact with (we aren’t able to see the source of all donations, so some could be included, although we were able to see the source of – and thus strip out – the bulk of donations for which this applies).
The following two charts show data just for December, the month in which we see significantly more donation activity.
Comparison to Charity Navigator and GuideStar
As shown above, donations through our website were substantially higher in 2010 than in 2009. We used the public Network for Good tickers for Charity Navigator and GuideStar to compare our growth (and level) to theirs, to see how much of our growth can be attributed to GiveWell’s improvement in particular vs. more interest in online charity evaluators / online giving in general. (We have confirmed that we are looking at the right data with Charity Navigator; we are still waiting for confirmation from GuideStar, but the numbers we got from this ticker match almost exactly with the independent numbers they sent us for 2010.)
It’s also worth noting the levels of (not just changes in) these figures. Though Charity Navigator had about 30x as many donations as GiveWell in 2010, and GuideStar about 9x as many, the average donation for GiveWell was over 4x higher (~$450 for GiveWell; closer to $100 for each of the other two). Net result: the total online donations for GiveWell were about 13% of those for Charity Navigator and 42% of those for GuideStar.
There are many possible interpretations of these numbers. One could argue that GuideStar’s and Charity Navigator’s numbers understate the actual impact, more so than they do for GiveWell, because people are more likely to use those sites as only one step in the process and end up giving via the charity’s own website. One could also argue the reverse: available information implies that more people are using GiveWell’s research to make large donations, and people often prefer to give these donations using donor-advised funds and checks rather than online.
One thing we do feel is the case, however, is that a donation given through GiveWell likely represents a much bigger impact – in terms of changing a donor’s actions from what they would have done otherwise – than a donation given through GuideStar or Charity Navigator. GiveWell recommends only 10 charities, while Charity Navigator lists over 1600 charities with its highest four-star rating and GuideStar allows donors to give to just about any charity they’d like. GiveWell is designed to make a strong recommendation of where to give; the others are designed to help people make one final check on the charity they had already planned on supporting. So while donations through the website remain lower for GiveWell than for Charity Navigator and GuideStar, they are now in the same ballpark, and there is an argument that the overall impact of GiveWell is at least as high.
What changed between 2009 and 2010?
- Six-figure gifts: two individual donors, both of whom have been following GiveWell for a long time, each made six figure gifts in 2010. The two donations totaled around $460,000. These donors had both refrained from large gifts to recommended charities in the past; in 2009, their giving to our top charities totaled $10,000. On the other hand, we did not repeat the 2009 economic empowerment grant (funded by an anonymous donor). So we gained a net of about $200,000 on six-figure gifts.
- New donors: $387,585 came from donors who hadn’t given before, while we lost $118,793 via donors who had given in 2009 and did not repeat in 2010. So the net from donor turnover was about +$270,000. This figure is about the same size as, though conceptually different from, the increase we saw in donations through the website.
We break down the 2009-2010 changes in more detail at the bottom of this post.
Our web traffic roughly doubled in 2010 vs. 2009. The following shows web traffic by source.
- The largest driver of growth in web traffic in 2010 was increased organic (i.e., non-AdWords) search traffic.
- The charts include data from January 2011 and the recent dips are a function of normal seasonality — i.e., we have more traffic during December’s giving season.
Money moved vs operating expenses over time
The following chart shows GiveWell’s total money moved each year (2007-2010) relative to our operating expenses. A major question for GiveWell is whether the value of our research is worth the cost of producing it. Money moved has continued to grow significantly relative to operating expenses.
Note that the above chart reports lower 2009 and higher 2008 money moved figures than our 2010-vs-2009 comparison above and our previous report. This is due to the way we reported data in 2008 and 2009. We spent a significant portion of 2008 soliciting funds for organizations we intended to recommend in the future. We ultimately made these recommendations in mid-2009. The funds were actually donated in 2009 but were committed in 2008. We believe that counting the funds when they were committed provides the most accurate picture of changes in GiveWell’s influence relative to operating expenses over time (though counting the funds when they were given makes it easier for us to track what changed between 2009 and 2010).
What we count as “money moved”:
- Donations made to top charities directly through our website. Though these donations go directly to top charities, we are able to track them and verify that they went through our website. (Example: VillageReach donate page)
- Donations that our recommended charities report back to us as being attributable to GiveWell (we have a high standard for this – we count only cases where (a) the donor explicitly stated that their donation was on the strength of GiveWell’s recommendation or (b) the donor gave to Nurse-Family Partnership and stated that they heard about it from a Nicholas Kristof column; Mr. Kristof has informed us that he included NFP in the column on our recommendation).
- Donations that donors report to us (informally or using our donation report form) as donations that they made on the strength of our recommendation. We cross-reference our data with recommended charities’ data, when necessary, to eliminate double-counting.
- Donations made directly to GiveWell and earmarked for re-granting. We count donations made and restricted in year X, and then granted in year Y, as “money moved” for year X, not year Y.
More details on what changed between 2009 and 2010:
- Two individual donors, both of whom have been following GiveWell for a long time, each made six figure gifts in 2010. The two donations totaled around $460,000. These donors had both refrained from large gifts to recommended charities in the past; in 2009, their giving to our top charities totaled $10,000. On the other hand, we did not repeat the 2009 economic empowerment grant (funded by an anonymous donor). So we gained a net of about $200,000 on six-figure gifts.
- $387,585 came from donors who hadn’t given before, while we lost $118,793 via donors who had given in 2009 and did not repeat in 2010. So the net from donor turnover was about +$270,000. Some additional context on these numbers:
- 13 new large donors, giving between $8-30k each, accounted for $163,449 of the $387,585 in gains. The rest came from nearly 500 smaller donors ($6k and under).
- Of the $387,585 in new donations, we can (using a combination of web analytics and optional “Where did you hear about us?” surveys) attribute $56,672 to organic search (i.e., people searching for things like “best charities” – not for GiveWell itself – and not including Google AdWords); $30,903 to the outreach of Peter Singer; $19,691 to “word of mouth”; and $18,517 to Google AdWords.
- We do not know the source of the other ~$195k in new donations. We can guess at what the distribution looks like using our survey data. Of the 2010 users who responded to surveys about where they heard about us, 13% found us via search, 23% found us through media, 25% found us through word of mouth or links from other sites, and 38% found us through the outreach of Peter Singer.
- Of the ~$118k in lost donations, $73k can be attributed to donors who simply gave early in 2011 instead of late in 2010. The remaining ~$45k comes from donations of $5,000 and under, and the fact that we lost this many small donors is a source of some concern.
- Donors who gave in both 2009 and 2010 gave less in 2010: $411,787 vs. $354,900 for a net loss of about $57,000. Three very large donors lowered their donations by a total of over $100k, for reasons that we believe to be related to financial circumstances, while another donor increased his by $30k; the rest of the changes netted out to a slight gain.
- Donations to GiveWell, earmarked for regranting, fell by a net of $38,326. About $46,000 of this loss represented donors who switched over to giving to our top charities, or to GiveWell unrestricted, instead of giving to GiveWell restricted. There was another ~$29,000 in losses, $9,000 of which came from one donor and the rest of which came largely from people who had donated immediately after our 2007 launch and not returned, offset by ~$36,000 in gains, ~$34,000 of which came from 3 donors.
- We had $238,988 in donations that we weren’t able to attribute to specific donors (donations that went through Network for Good plus donations reported to us by charities as coming on the strength of GiveWell’s recommendation), up from $56,200 in 2010, for a net gain of about $180,000 in this “mystery” category.