This very thoughtful post by Jonathan Mendez makes the point that publishers are sitting on piles of underutilized intent data, including search queries that are passed to a destination website when a user clicks on a search engine result. Mendez’ point is that by more effectively harnessing these kinds of data, publishers can extract more of the value they deserve from the ad ecosystem. He also makes the point that publishers enjoy a privacy advantage in exploiting this kind of data, to the extent it is “first party” data not subject to new privacy initiatives like do-not-track.
It’s not obvious that this is correct, and the question raises an important larger point about the new self-regulatory framework.
“Online behavioral advertising” is usually defined to involve the collection of user data across different websites over time. This appropriately excludes the collection and use of data by a website for the purpose of content and ads provided within the bounds of the website. Consumers are thought to understand this practice implicitly, and merely by visiting or not visiting a site, make their privacy choices.
On the other hand, a search query that is passed from Google to a website involves intent that was, by definition, expressed by the user off of the destination website. I haven’t seen any indication that users generally understand that queries are communicated in this way (it doesn’t typically show up in the URL itself). It’s even less likely that they understand that their search queries may be used for ad targeting (outside of Google).
The answer to this question is important for two reasons: First, if any publisher is engaged in cross-site behavioral advertising by virtue of onsite ad targeting based on Google search queries, then they would need to provide the same enhanced notice and choice of that activity as any other tracking company.
The answer also is important as a matter of precedent. To date, Google has not leveraged search queries to target ads across sites in Google’s ad network. But there’s little doubt that if (or when?) Google does so, regulators would consider this “behavioral advertising” subject to enhanced notice and choice. Would it be consistent to treat the same Google search queries differently when used by individual websites to target ads?
The fact is, referred search queries are already widely collected and used by websites to target content and ads, and there’s no doubt that it enhances the consumer experience. Fitting this practice into the new framework won’t be clean or easy, but it should be addressed.
PS We’re currently testing a new version of TrackerBlock for Firefox that includes the option to strip out search query referrals. By default this will be off, but in the course of offering it we will hopefully learn something about how users regard the privacy implications.