Crowdsourcing Effective Advertising

It does not take a genius to work out that ads for strollers are likely to be wasted on parents with older children or that you shouldn't put a Starbucks ad in front of someone who doesn't drink coffee. But how can you tell who is really looking at your ad? This week Google publicized various innovations it is bringing to display advertising that are designed to generate a better match between what is displayed and what a web surfer might actually want to buy.

Google uses your browsing history to make a prediction as to what type of person you are. For instance, I received some emails about summer camps for my children and ads popped up in Gmail for other possibilities. And, indeed, my daughter went to one of those alternatives this summer. Put simply, being able to keep track of what appears to interest a consumer is important information in improving the efficiency of matches with advertisers.

But it is instructive to think about what other information could be brought to bear to improve advertising. Advertisers care not only about getting in front of the right consumer but also about getting those consumers to pay attention, click, and purchase. Ads that are annoying — ads that flash or move to get your attention, for instance — do not fulfil this goal. In fact, an annoying ad in front of a consumer who is your target could be explicitly counter-productive, causing her to think worse of you and better of a rival. This has prompted Ian Aryes and Barry Nalebuff to suggest that ads receive ratings. Some platforms already ask consumers whether they find an ad useful or not. But these are small in-roads. So how can we help reduce the annoyance factor and increase the effectiveness of ads?

An obvious idea comes to mind: why not use a consumer's social network to help in making ads more targeted and less annoying? Social networks already have mechanisms to assist their users in sharing recommendations with their friends. Many, including Facebook, have a "like" tag for members to indicate to their friends whether their comments or links are, well, "liked." This assists everyone in understanding what is valuable and improving what is being shared.

What if social networks had the same option for ads so as to exploit socially held knowledge? Members could be asked to indicate whether they think their friends would like an ad, resulting in those friends seeing more or less of it. Advertisers could tap into social knowledge about what is likely to be an effective ad; something that is currently hard to achieve.

Why would consumers bother to "help out" advertisers by actively evaluating ads? My hunch is that a social norm could evolve to achieve just that. The "social contract" could become: you should help us bring less annoying ads to your friends by actively evaluating ads for them and then, to reciprocate, your friends will do the same for you. Everyone will be better off as the annoyances of ads you see are reduced, and advertisers will pay more for ad space.

Of course, one might be concerned about whether advertisers and platforms might exploit this newfound effectiveness to throw more ads at us. That is a legitimate concern, but also an opportunity to help the social norm evolve: the platform could guarantee that members who evaluated ads for their friends saw fewer of them.

Using social information to improve ad effectiveness has until now been at best a passive instrument. Bringing active engagement to the process is surely something worth considering.

Joshua Gans is an economics professor at the Melbourne Business School, University of Melbourne. He is currently a visiting scholar at the Harvard University Department of Economics, and the author of the book Parentonomics.