Google has never been shy about the fact that it automatically puts results from some of its own services at the top of search results pages if these help answer a users’ query faster. If you type in an address, a map from Google Maps might give you the instant answer you need.
But how often does Google do this, does it always result in superior results, and what effect does it have on companies that offer rival services? Ben Edelman, an assistant professor at Harvard Business School, has tested what he calls “hard-coding” by the search engine – that is, where it overrides the normal algorithmic results to put a Google answer first – and concludes the practice is now common.
The types of Google-derived information that come out on top range from movies, maps and health information to something as detailed as a search for patents (who knew Google had a patent search engine?)
In one search, Edelman found that nearly half the screen real estate on a search results page was taken up with Google services (though this probably overstates the extent of the hard-coding: Google says when YouTube videos appear in a result it’s because they’ve been returned by a regular algorithmic search, not because they’ve been favoured over other videos.)
Does this matter? The thought that Google is directing more traffic to its own services makes potential competitors nervous – as Barry Diller complained to me after Google agreed to buy travel technology company ITA.
The reason: providing a direct answer to a search for something specific (such as an address or an airline fare) is often the starting point to further action by the user (finding a restaurant nearby or buying a ticket), and Google has positioned itself to be the beneficiary of that.
When I asked Amit Singhal, head of search quality, about this issue a few months back, this was his response:
Our tendency tends to be, when a [Google] product has become popular with its users, then it’s OK to show it.”
Of course, that leaves plenty of questions. Is it only a tendency, what does “popular” mean, and what’s the cause and effect here – do services only prove their popularity after they appear at the top of the search results? And does it matter if a rival service might actually have produced a better overall result for the user?
In its review of the specifics of the ITA deal, the Department of Justice is unlikely to look into broader questions like these, but that does not mean they will go away.
