Ben Edelman is an assistant professor at the Harvard Business School whose research explores the public and private forces shaping Internet architecture and business opportunities.
Much of his recent work has focused on search engine marketing and it was a study he authored on the effect of advertising disclosure labels on clickthrough rate which got our attention at the end of last year. His research was especially timely on the heels of Google’s change in disclosure label from “Sponsored Links” to “Ads”. And I had cited his study as part of the research The Search Agency conducted on the impact of Google’s change on clickthrough rate.
It took a few months, but we were able to catch up with Prof. Edelman to learn more about his research interests as well as his perspective on our study:
Q: What motivated you to research the impact of disclosure labels on search behavior?
It’s an open secret that users have trouble recognizing search engines’ advertisements. And perhaps that’s no surprise. Search advertisements appear adjacent to algorithmic results. Fonts match exactly. (Google used to present advertisements in a font size one notch smaller, but that’s no longer the case.) Where search advertisements have background colors, those colors are increasingly pale and easy to overlook — literally invisible on many screens from certain viewing angles and in certain conditions (e.g. laptops outside).
Meanwhile, the vague label “sponsored links” makes complete sense to industry professionals, but I always had the sense that regular people don’t know what it means. My mother, an English major, taught me never to write in the passive voice, so when I see a passive-voice label — presented to millions of users, day in and day out — I’m particularly skeptical.
Q: What are your thoughts on the study we developed? What are some of the pros and cons of retrospective analyses vs. traditional empirical research?
I like your study. Of course you’re limited to the specific campaigns you’re working on, but there’s no reason to think your campaigns are unrepresentative.
Your study gives you across-the-board results, summed across all users. In contrast, the approach Duncan Gilchrist and I used lets us see which users are most influenced by changes in advertisement labels. In particular, we find that vulnerable users — those with low income, low education, and little online experience — benefit most from a change to the alternative label “paid advertisement.”
Q: There has been a lot of industry conjecture as to why Google made the switch from “Sponsored Links” to “Ads.” In your opinion, what prompted the change?
I don’t know. The best way to find out would be to ask Google, but Google hasn’t been forthcoming.
I suspect I’m not the only person who flagged shortfalls with the label “sponsored links.” Probably some regulators — perhaps state attorneys general — had the same concern.
Q: You have been somewhat critical of Google for its failure to satisfy the FTC’s call for “clear and conspicuous disclosures.” How could Google improve transparency for both searchers and advertisers?
First, you’re exactly right about the FTC’s requirement that disclosures be both “clear” and “conspicuous.” Those are two separate requirements, and both must be satisfied.
For US users, I believe “ads” is clear. Americans know what “ads” means.
Though that may not be the case for all users in all other English-speaking countries; in some age groups in some countries, I’m told the shorthand “ads” isn’t widely used.
But I don’t think “ads” is conspicuous. Last fall I looked at shortfalls in the “ads” label: I showed how the label “Ads” is so tiny it substantially fits inside an “o” in “Google”; it’s the smallest text anywhere on Google’s results page; it’s all the way in the far corner, which makes it particularly easy to overlook. At least “sponsored links” was larger — more letters, hence harder to overlook.
So what should Google do? Google needs a disclosure that is both clear and conspicuous. “Paid advertisement” — the label the FTC has required in other media for decades — would be the natural choice.
Q: Are you going to conduct further research on Google’s switch? If so, what information are you looking to collect?
In the most recent draft of my paper on advertisement labeling, I compared “sponsored links”, “paid advertisements”, and “ads”. I found that “sponsored links” and “ads” are statistically indistinguishable — very close in user response. “Paid advertisement,” meanwhile, yields fewer clicks but more accurate user recall of how many ads users saw and clicked.
I’m also thinking about other areas where Google advertisement labels fall importantly short. For example, Google now shows some advertisement content within algorithmic results, e.g. the yellow tags appearing with certain Places listings. That seems needlessly confusing to me: Under Google’s new approach, advertisements are at page top and right, and sometimes at bottom-left too. This gets to be altogether too complicated to explain to reasonable users. Google should return to its historic practice of keeping advertisements and editorial content firmly separated.
Q: Considering the increase in ad format sophistication (sitelinks, product feeds, ratings, maps, etc.),do you feel that even changing “Ads” to “Paid Advertisements” is a significant enough change to help searchers distinguish the sponsored from organic content?
I’m not sure it’s sufficient. But it would be a straightforward step in the right direction. Here’s hoping!