A recent study of Google search results by Tim Wu (of net neutrality fame) and others found that the search giant might be skewing its search returns to favor certain businesses. It's hard to know what's surprising about that. Google is after all a business, not a non-profit. Its business model is predicated on simultaneously providing users with a "good search experience" - so they become a go-to internet information portal - and providing businesses with opportunities to display their wares and their ads, thereby generating revenue for the company. A delicate balance needs to be maintained: giving the user (consumer) the sense that they are getting to the information they need, ensuring their customers (businesses) that they are generating profitable traffic and, of course, keeping their own business profitable and their shareholders happy.
A search engine is like a really great hotel concierge. Upon check-in, she already knows a lot about you. In fact, she may be the only concierge you've ever had. She knows all sorts of personal facts about you as well as your likes and dislikes -- some freely admitted and others inferred from your online activities. She also keeps meticulous records which of her recommendations has panned out. Even more, she knows similar things about all kinds of people, some like you and some not. You indicate to her with just a few words what you want to find or learn about and she presents you with a long list of options, ranked according to what she thinks you'll like best, based on what has succeeded before, for you and for many other people like you, but also according to what the world of options that seem relevant to this query have told her. Even more, some of them may have given her a little bonus to help steer customers their way. Ultimately, you as a stranger and naïf in this great world of possibilities trust her choices and you continue to go back to her and she continues to advise you, learning more and more about you in this kind of echo chamber of action and recommendation. How the decisions of relevance and the degree of relevance are determined make up the "secret sauce" of the concierge.
The skill of the concierge is then measured via the user experience. If you stop asking for her help, she's out of a job. If eyeballs, thumbs and fingers start to migrate away from the search engine, the SE business model falls apart. An interesting piece of this is the huge "information asymmetry" that is a part of web experience. We only know that part of the web that the search engines return to us. Users are at the mercy of the search engine with no way to judge whether the process was useful except for whether they found something that seemed "good enough". How could the user possibly know if there is something better out there? We rely on the big players in the search industry to be smart enough to design algorithms to help us find our information prey, but maybe they've missed something. Or worse, maybe they purposely aren't telling us something. To adapt an old joke, we ask where the keys are and the search engine decides which lamppost to turn on and where to point the light.
The work of Wu, et al. exposes some of the aspects of this darkened experience. The first is that it makes clear there is nothing objective or agnostic about the process of "search." A more interesting finding is around the conclusion that the results indicate users want more "diversity" in their search returns. Given the context of the experiment (supported by one marketing/advertising platform in combat with another), "diversity" here is code for the ability for other companies to have a stronger (or at least present) contribution to the Google search algorithm. That is, users were happier - or found more things that they liked - when given additional information or context in which to consider the search returns. This perhaps has implications not only for the antitrust actions that Google is facing (the related impetus of the reported research), but more generally might provide a lesson learned for the search industry.
For example, a broad conclusion is that more information for the searcher is better.
But more generally, what the Wu study might be showing us is that users appreciate the ability to engage more actively with the search process - i.e., not just take the results as is, but perhaps do a little information aggregation on their own. To that end, a more general approach that the search industry might consider is to pull back the curtain just a little on their process. Why not give users more agency? Why not give the user the ability to modulate the inputs to the search engine - like dials on dashboard - to enable the user to effectively simulate being someone else. That is, searching as if he or she was another person with different search characteristics, and in doing so, present the user with a presumably different view of the web? Effectively, let the user describe a "search avatar" to experience a different kind of search and also to affect the way in which the searching takes place. To date, we rely exclusively on the search companies to figure out what it is we want to get at. We can only change the way we explore the web through our actions of exploring the web. This is very inefficient. Why not let the user change the rules of the game - at least a little? Why not turn the ability to search into the ability to explore?