The other day a programmer I work very close with and I were discussing socialization on the web. It spanned everything from recommendation engines to crowd sourcing to search, which, of course landed us on the difference between algorithmic results (traditional relevance search) versus networked crowd-sourcing (voting).
The other day, Google announced it would be adding the “+1” feature to it’s suite, allowing members to effectively vote up links they like. Wired Magazine carried a brief overview which then spawned this response as I thought through the implications of such a move by the search giant as I think it’ll basically fall as one more piece of a very very complex algorithm that Google will continue to guard secretly.
The problem with “like” has always been there’s a difference between short term excitement and long term usefullness. some of the most “liked” FB pages are not necessarily the most relevant or useful when trying to research something. Plus, liking something is a very low interaction with it. There’s no barrier to use at all. “Liking” pages is as much out of impulse as anything else, not that much different than “friending” was on MySpace or “Fanning” was on iLike or even “Following” is on Twitter. Just because a page is “liked” does not necessarily mean my friend is actually recommending it to me. Nor, in many a case, does the fact a page is not “Liked” reduce significantly the likelihood my friend wouldn’t recommend it to me if I asked.
An example is people who are a fan of the roaming gnome of Travolicity might “like” Travolicity, but if I inquire to them who they use to book travel and lodging they very well could come back with some other companies and, in one case of someone I know who adores the campy commercials, has never booked anything through them despite being a frequent flyer. Another example is if I needed a mechanic, I doubt relying on the “Likes” of my friends would do much good, as there is nothing sexy about having STS or Pep Boys or your local garage “Liked.” However, those kind of mundane and seemingly trivial pieces of information are just the kind of thing search results or other kinds of interactions.
That is, in no way, meant to defend the “click” on a search result. How many times have you clicked on a result at the top of the page only to discover once reaching the destination it wasn’t really what you wanted. Google registers the click, knows that your search entry and the click match up and gives your click weight toward that link, in theory. In reality, it is a little more complex, but the gist is that erroneous clicks, ones used to investigate the usefulness of the result on links that may not be the final necessary result, are still given weight.
I think this is Google’s way of blending the more public crowd-sourcing of your network (and potentially, the aggregate of all users) with that of their current relevance algorithm to further refine results and make them seem more relevant. It will span the way people use search for some kinds of recommendations and their network for others in a way that could become extraordinary powerful because it takes into account a greater number of touchpoints than either system currently entails.
It also, as the article mentions, is a way for Google to leverage more of the personal information it has on you and your network in direct competition to Facebook (and Twitter) in order to stay relevant for advertisers and moreso investors. Remember, Google’s been an advertiser darling for ages but they are facing increased pressure on targeted ads by places such as Facebook. Performance on ad revenues and public mindshare both effect investors and with that being said, as the push for Facebook to go public comes ever closer the possibility of Googles still stratospheric stock coming down to reality looms ever larger.
Google needs to produce another successful product, not just in response to the outrageous failure of Buzz, but as a way to sure up their advertising income beyond the traditional revenue generators associated with those ads. If this works, it not only helps provide remedies to both of those issues, it helps define some parameters for true social search. If not, it may just prove to be an idea either too far ahead of it’s time or not executed in the most efficient manner because at some point in the future some kind of blended search will come to the forefront. Personally, I think there two layered components that are both fighting for their place in the search / ad world: the crowd-sourced voting concept being one and geo-location being the other.