Gifts for donors

Sometimes nonprofits provide donors with gifts as a way of thanking them for their donation. Does this work? Several experiments have tested this practice.

A face to face fundraising experiment among visitors to a national park in Costa Rica (Alpizar 2008) found that giving potential donors a small gift increased the likelihood of them donating, but lowered the average amount donated. The experimenters approached tourists just after visiting the main attraction of the park and gave some of them a handcrafted refrigerator magnet worth about $3. About 56% of those who got the magnet donated money, as opposed to 48% of those who did not get a magnet. They gave less money, however, and average of $4.56 compared to $5.09 for those who did not get a magnet. Apparently a feeling of obligation led some people to give a small donation in the gift condition when they would have made no donation without the gift.

An experiment with direct mail also tried the effect of gifts (Falk 2007). A researcher worked with a large international children’s assistance organization based in Zurich, Switzerland, and divided a Christmas mailing of 10,000 letters into three groups. One group received no gift, the second received a postcard and an envelope, and the third group received four postcards and four envelopes. The postcards were pictures drawn by children in Bangladesh whom the charity assisted. Only 12% of the people solicited replied to the letter with no gift, but 14% replied when there was a small gift and 21% replied when there was a large gift. They also gave larger amounts: an average of 7.56 francs with no gift, 8.37 francs with a small gift, and 12.21 francs with a large gift. Overall, the large gift mailing was much more successful, bringing in 40,877 francs as opposed to 27,106 from the small gift mailing and 24,673 from the no gift mailing.

[table id=1 /]

A third experiment (Landry et al. 2010) used face to face appeals with two types of donors to a university research center – a “warm list” of people who had previously donated to the charity, and a “cold list” of people who had not previously been asked. They used three experimental treatments: a third of the people were just asked for money, a third were given a small gift (a bookmark) and a third were given a large gift (a copy of the book Freakonomics).  In the absence of a gift, warm-list households were much more likely to give money and gave larger amounts. A gift, however, evened the odds, so that warm-list and cold-list households were about equally likely to make a donation. The gifts worked on cold-list households but had little effect on the warm-list households, raising the cold list households to the same level of generosity as the warm list households but having little additional effect on the warm-list households.

  • In the first experiment, gifts made it likely that people will send in a donation, but the extra donations were small.
  • In the second experiment, gifts increased the likelihood of donating and the amount donated, and a large gift had a stronger effect than a small one.
  • In the third experiment, gifts worked but only on those people who had not already donated.
  • Gifts seem to work and often pay for themselves, especially on first-time donors.

If giving a gift seems like a plausible strategy for your organization, try your own experiment – give a gift to some random sample of prospective donors and see if it increases the amount of money you collect, compared to the random sample of donors who did not receive a gift.


Alpizar, F., Carlsson, F., & Johansson-Stenman, O. (2008). Anonymity, reciprocity, and conformity:        Evidence from voluntary contributions to a national park in Costa Rica. Journal of Public              Economics92(5), 1047-1060.

Falk, A. (2007). Gift exchange in the field. Econometrica75(5), 1501-1511.

Landry, C. E., Lange, A., List, J. A., Price, M. K., & Rupp, N. G. (2010). Is a donor in hand better than two in         the bush? evidence from a natural field experiment. The American Economic Review100(3), 958-983.