Facebook Graph Search

Google co-founder Sergey Brin warned last year of a cyber dream gone bad. With the announcement of new search engine Graph Search, Facebook has edged one step closer to realising Brin's nightmare. There are "very powerful forces that have lined up against the open internet on all sides and around the world", the billionaire computer scientist said, citing "walled gardens" (cough, splutter, Facebook) as major threats to the values on which the internet was founded. He honed in on Zuckerberg's social network:
"You have to play by their rules, which are really restrictive … the reason that we were able to develop a search engine is, the web was so open."
The question has never been whether Facebook wanted to become the dominant force in controlling the online experience -- but how it would go about achieving it. There's been talk of Facebook web browsers and phones but in an age of cross-platform and mobile/app-based access they're becoming steadily less relevant; new ways to communicate will change smartphones as we know them. But the need to search the internet? That's here for good. Graph Search, an extension of Facebook's Open Graph, was announced to the tune of fawning headlines, from "Facebook's Graph Search Could Be Its Greatest Innovation" to more dramatic turns of phrase like "Facebook Just Declared War on Google: Meet Your New Search Engine". Google, the world's most popular search engine, was built by code junkies monetising the quickest and most efficient ways to deliver relevant search content under the assumption users will reward the service by returning. The result is the ultimate "bounce" site: visit, click, leave, return. Then there's Facebook, the former online pick-up joint for randy college students turned internet superpower. Its key advantage is not that it's the world's most popular website -- though that doesn't hurt -- but the amount of time users spend on it, a metric Facebook has long dominated. You don't have to be a fly on the wall at Facebook HQ -- where a new sign touting Graph Search has replaced its ubiquitous "like" logo -- to hear the boardroom rhetoric: users will never have to leave. Experiments in streaming videos, text and video chat, Facebook email and even reading the news have long hinted at what was coming. It's primitive, but Graph Search is a glimpse into the future of search engines: the capacity to sort results from information submitted by people users know and trust rather than invisible algorithms cooked up by Silicon Valley nerds. The browser of the future is not a browser at all -- it's a social network tailoring the web for individuals. I've been test driving Graph Search since last Saturday, accepted as a beta tester (anyone can apply) in the weeks/months before it will become de rigeur for hundreds of millions of users. Like any new system of searching or browsing -- from the early days of Geocities' branch-like community structure to Yahoo categories, Google keywords and Twitter hashtags -- Graph Search will be of limited value until its syntax becomes second nature. We know that entering a noun prefaced with "what is" is a fast way to find relevant information on Google; the same approach won't work as well on Graph Search. Graph Search asks users to enter "simple phrases" in order to return "personalised results [that] show details just for you". It plays to Facebook's strengths: people, places, pictures and recreations. The formula will be the subject of SEO detective work for some time, but searching a simple phrase like "Chinese restaurants" offers obvious indicators. Geographic proximity, "check-ins" (people who have logged in remotely with mobile devices), photographs and "likes" from friends are weighed against each other, the latter appearing to be the most influential variable. Small businesses will be scrambling to secure a Facebook page to ensure their place in search results, if they haven't already. And an army of SEO consultants will be offering tricks of the trade for climbing higher on results pages. Market analysts will pore over Graph Search to find what products and services appeal to certain demographics. For example, if Crikey's subscription department run a Facebook campaign offering a free DVD with membership, Graph Search can reveal what brands users prefer. Take a choice between Breaking Bad and Boardwalk Empire. If I search for "people who like crikey.com.au and who like Breaking Bad" the engine returns more than 1000 results/people; the same search for Boardwalk Empire returns more than 100. Crikey's readership -- at least those on Facebook -- prefer their crime dramas to revolve around meth-cooking suburbanites than prohibition-era gangsters. There are more than 40 ways to refine the search. For example: