Ten Thoughts On Facebook’s Graph Search

It’s probably become clear from my recent string of posts and tweets that I don’t hold Facebook in particularly high regard either as a tool or as a “tech” company. Basically, I think it is an ok but annnoying means of keeping up with friends, but nothing close to the sorts of revelatory software/hardware advances made by the likes of Apple or Google.

Today’s self-described big announcement from Facebook turned out to be a tweak to how desktop-formatted in US English produces its search results. Inelegantly called Graph Search, this new tool is a natural language means of trying to find, for example, “Indian restaurants liked by friends from India,” or “shoe stores with good reviews nearby.” It gets this data from posts (songs listened to, places visited, wall posts) made by friends or strangers which are viewable to you, otherwise known as the Open Graph (hence the new tool’s name). CEO Mark Zuckerberg carefully described this as a “beta of a V1,” meaning that sloppiness is expected.

Here are ten quick thoughts on Graph Search.

1. It’s a bit like a dating site.

Dating in fact was mentioned as a use case. Users of OK Cupid and apps of its ilk are likely familiar with the granular filters that can be used to find, for example, a “single man 25-30 who lives in Greater Chicago and is into LTR.” Graph Search introduces this granularity to the data on the Open Graph.

2. The average Facebook profile will need to add more information and make more of its information public in order for Graph Search to really work.

Google Search works because the object of its search – the open Web – is, well, open, and able to be parsed. Facebook is an odd walled garden in which much information is withheld from the majority of users. This will need to change if Graph Search is going to be an accurate, robust tool.

3. Facebook must fix its mobile apps (especially on Android) if Graph Search is really going to compete with Yelp, Foursquare, or even Google Now or Siri.

Facebook’s mobile apps are an embarrassment. Slow, unstable, bloated with features, and encumbered by confusing UI, they would have instantly failed if not for the huge user base that Facebook had accumulated at the tail-end of the desktop computing era. Would-be competitors like Yelp and Foursquare provide much better mobile app experiences. Platform-ingrained tools like Siri and Google Now are only going to get better at contextualizing information and figuring out who you are and what you want to do. For Facebook to compete, they have to make better mobile software, plain and simple. The potential use cases for Graph Search are almost entirely mobile, and would need matching software to be realized.

4. This is basically Search Plus Your World MK II.

Google rolled out Search Plus Your World last year, which used bits of information from your Google+ pages to personalize Web search. It hasn’t set the world on fire, which could be attributable to murky levels of engagement from Google+ users, but it created the blueprint for this type of move toward personal search.

5. It still isn’t clear if Facebook users use Facebook for much other than simple status checking and chatting.

Facebook has ambitions to be a commercial and technological juggernaut. But at present, Facebook is a lot like a bar – a place to socialize, but not quite the right setting for being pitched ads, getting expert answers to questions, or buying retail items. Maybe this will change. But it’s questionable if the average Facebook user wants much more out of the site than quick blurbs from their friends which they can like or comment on.

6. This makes Facebook at the very least searchable – it wasn’t searchable at all before.

Facebook is one of the most unsearchable sites on the Web. You can find friends and sometimes strangers, but search results are unpredictable, messy, and confusing, especially if searching for someone with a common name. Graph Search brings it up to basic usability, which should be applauded as a relief rather than a revolution.

7. If Graph Search works, it will mark the transition of Facebook from website to app/service.

Facebook is still a website. Its best experience comes from desktop use, an obvious legacy of its roots in a time before apps or iOS/Android. As mentioned above, its apps are passable at best. But creating a usable Graph Search will require a robust mobile experience that can get people accurate information on the fly. Doing this successfully would move Facebook away from being a static location (a website that has barely changed in nine years of existence, despite all the Timeline/News Feed rebranding) to being a dynamic service.

8. This probably isn’t a threat to Google or Yelp or LinkedIn.

I leave Foursquare out of this because I think Foursquare is in deep trouble of its own making and won’t last much longer. In any case, the use case for personalized Web search remains unclear, and I’m not sure that Graph Search fully delineates it – as a beta confined to desktop use in the US English-speaking world, it could be a while before it makes real impact on mobile or enacts the site-to-app change I just talked about. In that time-frame, we could easily see the entire mobile landscape shift, either to dynamic platform aggregation services with a big head-start (Siri, Google Now) or other apps which chip away at the amount of time spent on Facebook (Snapchat, Tumblr in particular). Facebook’s current (and perhaps transient) lack of seriousness about iOS and Android could cause trouble for the viability of Graph Search.

9. Graph Search/Bing integration is, along with the Xbox, now the most interesting aspect of Microsoft’s business

With the Windows brand fighting for relevance, Microsoft needs new outlets for revenue and user intake. Graph Search, if it can’t fully answer the question posed by the user, will default to Bing search results. This is a boost to Bing and a slight ding of Google, but it seems like just another instance of platform wars – Siri and Google Now also return search engine results when the question stumps them.

10. “Big announcements” aren’t what they used to be.

Six years ago this month, Steve Jobs made the “big announcement” of the original iPhone, a device whose importance can’t possibly be overstated. But the important thing to remember about Apple’s presentation – something that sets it apart from make would-be pretenders – is that Jobs had a real, shippable product in his hands, one that was definitely going to be released in a short amount of time and which wasn’t a “beta” or “draft” or far-off dream. Graph Search still seems like something not yet born, even though it was announced – sort of like Google’s still-MIA Nexus 4 charging orb. That’s not to say it’s vaporware (it’s not) but the gravity of its announcement is eroded by all the qualifiers around it (“beta,” “V1.0”).

-The ScreenGrab Team


2 responses

  1. […] You may say “so what?,” but there’s more. Via its own push notifications and widgets, Friday can suggest what you should or at least are likely to do next – such as calling a specific person, playing a specific song, or visiting a specific place. It keeps track of your favorite contacts and checkin spots to do this. You can also do plain language search to have the app return answers to queries like “who did I call at this time two weeks ago?” or “what song have I listened to the most.” ¬†Granted, some of these features are admittedly in “beta,” but the app’s overall attempt at gathering, remembering, contextualizing, and suggesting information is one of the more ambitious mobile app projects I’ve seen, and a major part of why I was so unimpressed by Facebook’s somewhat belated Graph Search project. […]

  2. […] too. With the gradual pollution of the News Feed and the concomitant rise of Graph Search (even if overrated), there may be an opportunity for individual profiles to shine […]

%d bloggers like this: