My Facebook Graph Search Notes

All you need to know about Facebook Graph Search

Facebook chief executive 010

These are the notes I made based on last night’s research, and the best links I’ve seen shared so far. All the “Tips” posts seem a little premature given the limited Beta roll out.

Facebook’s recommendation to Brands

  1. Make sure your Page, Place or App information is complete and up to date
  2. Strengthen your connections.

So — business as usual there, then. Get your Page in order, grow your fans.

Facebook’s “Introducing Graph Search” (and waiting list sign-up)

Implicitly, this is a better way to search Facebook. You want to find that girl you met last night at that guy’s party and you can’t remember her name? Want to make a list of all your friends who live in that town you’re visiting? Here you go.

Primarily, this will improve the Facebook user experience.

NB: The beta is rolling out in US only.

The Verge’s liveblog (with photos)

NB: The whole top bar title area becomes a search bar?


Chad Wittman’s “Is This The Facebook Search We’ve Been Waiting For?”

A late, but useful addition to this list. Chad points out:

Businesses with a physical location, AKA local businesses, will benefit the most from Graph Search. The second most benefitted businesses will be ecommerce. This should hold true at least through the initial phases of Graph Search. These businesses have the easiest input signals into the Graph Search algorithm, while also possessing clear-cut opportunities to obtain sales.

But points out many of the potential flies in the ointment, notably that the problem Graph Search is being hired to solve isn’t necessarily clear, or well-recognised.

Jesse Brown’s “Facebook’s B.S.-powered search engine”

Brown points out that the quality of the behavioural and surrendered data that Facebook is relying on is patchy at best. This will feed back into user experience.

Graph Search is only as good as the information we give to Facebook. And my Facebook information is garbage.

My profile does not include my employers or my alma mater. I don’t “check-in” when I visit a location, nor do I rate restaurants or movies on Facebook. I don’t “like” things because I like them, I like them when I’m trying to help my friends promote something, or to make a cheeky joke.

Venture Beat’s “Facebook stock closes down at $30.10 after announcing Graph Search”

Stock is down after rising on expectations of announcement.

Investors, interested in a new way to make money off of Facebook, pumped the shares further up to a high of $32, but the lack of information about a monetization strategy or advertising in search caused the stock to remain in the red. It closed today at $30.10 a share, down 2.74 percent.

Just as likely to be because investors “buy the rumours and sell the news”. I suspect the lack of “monetization strategy” could be a red herring here.

Comscore’s “What History Tells Us About Facebook’s Potential as a Search Engine” (June 2010)

Early indications that Facebook users were already heading in this direction

the fact that we are seeing the first real signs of a burgeoning “traditional” search experience bodes well for the future potential of Facebook as a search engine. I anticipate that we will see this type of consumer behavior evolve along the same lines of traditional search as more dollars flow towards social media.

My take outs

  • This is primarily about improving user experience. Facebook’s search has long been its Achilles heel. Despite being more or less fit for purpose (it readily identifies the John Smith I’m most likely to know, rather than the most famous John Smith) it has always been a lacklustre experience. And yet, anecdotal evidence from Facebook suggests that users are treating it like any other search box; as a means to navigate the wider web. Furthermore, there are other kinds of search (“which of my friends live in London?”) that have been impossible to date.
  • Graph Search searches behavioural data (listens, likes, checkins) and surrendered data (profile information). Not all these data will be good – or rather, there will be a spectrum of reliability. Spotify Listens are probably good data (unless, like me, you share an account with a whole household.) Restaurant checkins may be heavily biased towards those offering checkin deals, and Page Likes to the biggest advertisers.
  • Graph Search may not lend itself to all brands I doubt that “what soft drinks are most popular among my friends” will be heavy volume, whereas “what restaurants are good in Dublin” could be.
  • User experience is key to the development of this search. User satisfaction with results will determine what kinds of search become popular. Facebook is particularly good at responding to behavioural data, and I’d expect to see the search become optimised for these (if only in terms of typeahead prompts.) I’d expect to see some interesting differences between mobile and desktop usage.
  • The Page has become more important. Until this announcement, I had come to believe (with many others) that Pages might be in decline except as a means of injecting fan-endorsed stories (and ads) into users’ news feeds. The new search may well restore their strategic significance.
  • Open Graph objects are increasingly important. This trend continues. Brands and retailers must Open Graph-enable their owned spaces and e-commerce engines if they want to appear in search.
  • Feedback from search data is essential for brands to understand how to optimise their search results Google has a strong set of planning and feedback tools. We know search volumes, search rankings. It’s not immediately obvious that these will be available to Facebook advertisers in the short to medium term (or in the case of rankings – ever.)
  • Bing optimisation may have increased in priority. Bing results will (as ever) be included in the results; although (I assume) only when Graph Search fails, or as secondary material. However, the news looks good for Microsoft (and, indeed, $MSFT seems to up on the announcement.)

Please tell me what you think.