# Suggested Donations

One common practice among fundraisers is to suggest a giving amount to potential donors. Does this work?

The short answer is that it can work, but it’s tricky. The problem lies in where to set the suggested amount. If you set it low, more people respond but some of them give less money than they would have if you hadn’t suggested anything. If you set it high, some people rise to the challenge but many people don’t give at all. The decrease in overall response offsets the gain in average donation. This article first reviews a few experiments that didn’t work, then shows the results of several experiments where fundraisers did find just the right suggested amount that did result in increased donations.

In one unsuccessful field experiment, people who had attended a performance at a German opera house got fundraising letters asking them to support a youth music project connected to the opera house. The donors were randomized into three groups: one with no suggestion, one with a suggested donation of 100 Euros, and one with a suggested donation of 200 Euros. Both suggested amounts were above the average donation. In this case, people in the suggested donation groups responded by making the suggested donation, but fewer people responded than in the no suggestion group. The increase in average donation was balanced out by the decrease in the response rate, so the amount raised was the same (Adena, Huck & Rasul 2014).

A second field experiment involved a German website portal for charitable giving, and was a huge experiment: over 680,000 people visited the webpage during the experimental period and 23,000 of them made donations. Visitors to the website were randomly assigned either no suggested donation, a suggested donation of 10 Euros, a suggested donation of 20 Euros, or a suggested donation of 50 Euros. These amounts corresponded to the 25th percentile of all donations (10 Euros), the median (20 Euros), and the 75th percentile (50 Euros). There turned out to be no perfect spot for a suggested donation that would increase overall donations. When the suggested donation was low, more people gave but they gave lower amounts. When the suggested donation was high, fewer people gave but the average donation was higher. These factors averaged out, meaning that total donations were similar in all four groups (Altmann et al. 2014).

While these two experiments did not work, several others did. One (Alpizar 2008) involved asking money from visitors to a national park in Costa Rica. After visiting the main attraction of the park, tourists were approached by solicitors who asked for a donation. Sometimes the solicitors did not suggest a donation, sometimes they said that they had approached “tourists from many different countries and one of the most common donations has been [amount],” and they set the amount randomly at \$2, \$5, and \$10.

When the interviewers said nothing about a common donation, 47% of tourists gave money; this went up to 49% with a \$10 suggested donation, 50% with a \$5 suggested donation, and 61% with a suggested \$2 donation. Suggesting a donation brought the amount donated down, however, from \$6.00 when there was no suggestion to \$5.97 with a \$10 suggested donation, \$ 3.95 with a \$5 suggested donation, and \$3.61 with a \$2 suggested donation. Overall, suggesting no donation or suggesting \$10 raised the largest amount of money; suggesting only \$2 brought in donors but brought the average donation down so much that it raised less money than making no suggestion at all.

Fraser et al (1988) did a door to door fundraising campaign with 640 people living in Columbus Ohio, asking for contributions to the local branch of the Humane Society. They both set a large and a small suggested amount by stating, “We are asking people to contribute (\$20) to the Humane society. Would you like to make a contribution? (Even a penny will help).” In some conditions they asked for \$20, in some they added the phrase “Even a penny will help,” and in some conditions they did neither, asking only for people to contribute money to the Humane Society without specifying an amount. They found that adding “even a penny will help” by itself doubled the donation rate, but combining that with a \$20 contribution had little effect. Suggesting a \$20 donation reduced the number of people who gave but greatly increased the average amount given. The \$20 donation suggestion raised the largest amount overall:

Condition% who gaveAverage donationAverage revenue per donor
Control15%\$4.00\$0.59
Penny will help34%\$3.02\$1.01
Suggested \$2011%\$11.61\$1.25
Penny and \$2018%\$4.88\$0.89

A third study (Martin & Randal, 2008) looked at suggested amounts, this time in donation jars. They experimented using a public art gallery in New Zealand where admission was free but for which donations could be dropped into a transparent box in the foyer. They varied the amount and type of money displayed in the box when the gallery opened to send different signals. There was an empty box treatment, which sent the signal that most people do not give; a treatment where there were many 50 cent coins and a few dollar bills; a treatment where there were lots of \$5 bills, and a treatment where there were a few large bills, mostly \$10. They measured the total amount given each day for 52 days and came up with these results:

Condition% who gaveAverage donationTotal donations
Empty1.9%\$1.97\$205.25
50 cent coins3.4%\$1.69\$308.00
\$5 bills2.6%\$2.38\$316.65
\$10 bills2.3%\$2.39\$293.80

Again, having a small suggested donation increased donation rates. It decreased the average donation slightly, but more than made up for it in having more people donate, so the total earnings were much higher. The \$0.50 and \$5.00 donation worked about equally well, but a box containing mostly \$10 bills was too high of a suggested donation.

##### Conclusion:
• Giving donors suggested donation amounts influences their behavior – they will often give the suggested amount.
• A moderately high suggested donation seems to work best, but scientists don’t yet have a formula to determine what level this should be. If you adopt this strategy, you should experiment with different suggested donation levels until you find an amount that works.
• If you want to recruit new donors, a low initial suggested donation might be a good idea to recruit them onto your list. You can follow up later and ask for a repeat donation.

### References:

Adena, M., Huck, S., & Rasul, I. (2014). Charitable giving and nonbinding contribution-level suggestions:       Evidence from a field experiment.

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.

Altmann, S., Armin, F., Heidhues, P., & Jayaraman, R. (2014). Defaults and donations: Evidence from a   field experiment.

Fraser, C., Hite, R. E., & Sauer, P. L. (1988). Increasing contributions in solicitation campaigns: The use                of large and small anchorpoints. Journal of Consumer Research15(2), 284-287.Martin, R., & Randal, J. (2008). How is donation behaviour affected by the donations of others?. Journal of Economic       Behavior & Organization67(1), 228-238.