The unintended consequences of Wikipedia nudges

Nudges are one of the 'in things' at the moment, for business and policy. We have witnessed the rise of the Behavioural Insights Team (a.k.a. the Nudge Unit) in the U.K. and the Social and Behavioral Research Team in the U.S., and even the Australians (NSW Department of Premier and Cabinet) have a Behavioural Insights Community of Practice. But do nudges always work as intended?

Not according to a post earlier this week on the Misbehaving blog, by Jamie Kimmel. Of the three examples he provides, my favourite is the Wikipedia example:

Wikipedia began testing new ads on its website to bring in more revenue via donations. The central theme of these messages is around two facts: Wikipedia has 450 million users, and less than 1% of those readers donate to the site.

Problems arise, however, when we look at evidence from behavioral science around the use of normative statements. This one from the context of voting: 

"The fact that many citizens fail to vote is often cited to motivate others to vote. Psychological research on descriptive social norms suggests that emphasizing the opposite—that many do vote—would be a more effective message. ...Practically, the results suggest that voter mobilization efforts should emphasize high turnout…. More generally, our findings suggest that the common lamentation by the media and politicians regarding low participation may undermine turnout."

So it’s possible that, by pointing out how few donate, these Wikipedia messages may actually be counterproductive.

We don't have any evidence to support the failure of this nudge, but it does seem plausible that it isn't working as intended (and may actually be counter-productive). Quasi-rational people are affected by framing. If less than 1% of users donate to the site, that might still be millions of donors. Thus, it might be more effective to use a positive framing: "Join the 2 million others who have donated to Wikipedia this year!" or something like that.

Having said that, as you would expect it appears that Wikipedia is using some A/B testing to determine the most effective strategy for generating donations, according to this article:

Settling on the best format for ads is an ongoing challenge for the fundraisers at Wikipedia, who test different layouts throughout the year to see what gets the most engagement. Ads that garner most traction are then prioritized for the December campaign.

So, perhaps the negative framing is working? Or maybe they just haven't tested a positive framing yet. I guess we will find out next month when Wikipedia starts its annual fundraising campaign.