A search engine that needs instructions is doomed to fail. This has been my feeling since turning on Facebook’s new Graph Search. With all of their money and know how, Facebook couldn’t even make a dent in Google’s search empire.
I really felt that applying a real search option to Facebook’s social data would generate the type of results and platform that would usher in a new form of search, one that could bring real answers back for what the world is thinking. Facebook’s Graph Search is nothing but a fancy organizational tool for a user’s friends and their circles of friends.
Yet I can deal with that. What bothers me more is the stiffness of the search. Queries need to be structured in a certain way and the data they use is essentially the data that we mark down ourselves when building a Facebook Profile.
I tested out a simple query: “Which Friends of Mine are Jewish?” I was surprised that hardly any came back and of course the fact is I go through my friend list and can measure easily 90% that are of the Jewish faith.
So what went wrong? It seems that Facebook is not using action based data, like from stories, posts, and images that a person likes. They are in a sense filtering a user’s pre-picked interests and profile, data which is at the most very surface level.
I am hoping that guys (and girls) out in Palo Alto begin to dig down and make a truly sophisticated search engine based on more than just vertical matches, after all we expect a whole lot more.
Posted by David on January 27, 2013
All complex search engines, whether they are Google or Bing, and even shopping platforms like Amazon and eBay use search hierarchy as way to convey meaning. Ordered search hierarchy is the best way to tell and differentiate between useful results. We have grown used to seeing you tubes at a certain place for related keywords or Wikipedia’s dominating search queries that encompass named entities, so much so that when they are not there are brains tell us that something is off. Structured at their basic level create meaning and definition within a set of data. Shopping platforms like Amazon and eBay rely on ordered search results as well to give rich meaning to a particular search.
The greater the definition and clearer the parameters the more meaning there is behind any given search query. This is ultimately the idea behind Google’s Knowledge Graph as well as Amazon’s “Buy Box.” Usefulness in search creates higher CTR and a more valuable and informative search experience.
Ultimately Google and others know that the best search are those that are useful and essentially our role as search experts is to fit our information into a search engines set of parameters.
Posted by David on January 20, 2013
Below is the Google Trends for three of the current party leaders in the upcoming Israeli elections:
Naftali Bennett, Bibi Netanyahu, and Yair Lapid
If this trend continues, Naftali Bennett the current dark horse of the election will continue to surprise pundits!
Posted by David on January 2, 2013