or Can people tell you which causes they’ll support?
There are plenty of surprises in the causes people choose to support on eBay for Charity. While the most obvious causes (children, animal and cancer-related charities) tend to attract plenty of support, much depends on the exact nature of the fundraising message we show to eBay users.
Through eBay Give at Checkout, we’re able to show short fundraising messages to many thousands of people each day, inviting them to add a small donation to their purchase on eBay. Since launching in 2008, the feature has raised over £5m for UK charities. A couple of examples are below:
We’re obviously interested in whether people’s giving behaviour on eBay is similar to their expressed preferences, and giving behaviour in other contexts. To understand these relationships better, we looked first of all at causes by overall popularity, using NCVO and CAF’s UK Giving report.
We didn’t find huge surprises here: most of the popular causes tend to perform well on eBay. However they are a little broad and vague to be useful in helping us find and feature new causes people are likely to support.
Another obvious (though expensive) tactic is simply to ask people which causes they support. One doubt was raised though: how reliable will the information be? After all, as Karl Wilding of NCVO has highlighted, giving surveys are particularly vulnerable to “social desirability bias”, in which “individuals respond to survey questions in a way that makes them look good, avoids embarrassment or tells the interviewer what they want to hear”. Would this affect surveys concerning specific causes: for instance, are people likely to say they will support sophisticated international development charities, but when shopping on eBay they might prefer to rehome abandoned kittens? By working with research firm nfpSynergy, we were able to find out.
We chose a wide sample of causes and displayed each of them to more than 500,000 users on the eBay site, asking each to donate £1. Taking 1 as the average response rate within the sample, we indexed the causes by popularity among eBay users.
nfpSynergy then surveyed (online) a nationally representative sample of 1,000 adults. They asked the question, “If you had £1 to donate to charity, which of the following causes would you be most likely to donate to? Please select up to 3”. The wording of the cause and amount of the donation was identical in each exercise.
The causes are shown below, ranked by popularity in the survey shown by the pink bars. The blue bars represent the causes’ popularity at checkout amongst eBay users.
There appears to be a fairly broad correlation between the two. The least popular cause is the same on both eBay and in the survey, while the top two in the survey are both strong performers on eBay. On the other hand, there seem to be some real outliers – the third cause (supporting small, local charities), and the second last (helping a child in Kosovo). You can see this in more detail in the scatter diagram below.
While there is a trend line, very few points are close to it. The causes marked out in green perform much more strongly on eBay than in the survey, while those in pink do better in the survey.
What could explain the differences?
Some of the variation may be due to sample differences (frequent eBay shoppers versus the general population). Another possible explanation lies in the emotive component of the messages, which is most evident in the two furthest outliers.
- “Help small UK charities…” may make a lot of sense in principle (survey), but is perhaps too abstract to inspire an actual donation. How exactly would the donation help?
- “Give a child in Kosovo a pair of shoes…” might seem an odd cause to say you’d support in the survey, but conjures up a powerful and emotive image that can trigger a donation.
Some of the other outliers may derive from the complexity or controversy surrounding the issues. Feeding a malnourished baby sounds simple, obvious and helpful, whereas intervening in families with potentially more complicated needs (involving asylum seekers or child carers) perhaps less straightforward.
If we accept that explanation, there are some striking implications for fundraisers. First, potential donors may turn out be more responsive to simple, emotive imagery than they claim or indeed believe. Conversely, donors may overestimate their interest in discriminating between causes on rational grounds. In shaping fundraising campaigns or messages, fundraisers should not imagine that potential givers will make a rational judgment in response. There is plenty of scope for fundraisers – even those supporting somewhat obscure causes – to capture the attention and commitment of potential donors through the skilful use of imagery and language.