Why Graph Search will be More Impressive Than You Think
April 16, 2013
Five years from now, Graph Search will have won the search war. Not because it’s going to have better algorithms than Google, or have better reviews than Yelp or Foursquare; but because unlike those services Facebook is a platform and not a provider. The future value of Graph Search is only going to be as strong as the providers that utilize Open Graph.
A lot of attention is being paid to Graph Search, and the essential quality of “likes” and how both users and advertisers can utilize this information to improve the value they receive from Facebook. However, few of the critics of Graph Search have gone beyond the “likes” and looked into the entire framework on which Open Graph is built. Content providers such as Spotify, Fab. and Goodreads have done a fantastic job of using and contributing to Open Graph, creating the foundation on which Graph Search will win the next great search war: social aggregation and discovery. Advertisers have already had significant success and ROI with Open Graph Sponsored Stories, which in turn perpetuates the Open Graph architecture. Open Graph provides value to both the advertisers and consumers.
The future of Graph Search will have little to do with “likes.” Instead, it will be the aggregation of the endless number of other actions on the graph that will be indexed to produce a superior search engine. “Likes” have become far too passive to provide useful data. In the next few years, as Facebook attempts to tackle the massive project of integrating the entirety of Open Graph into Graph Search, the concept of “likes” will be close to irrelevant.
The Open Graph architecture breaks down a person’s presence on the internet into “query-able” bites of information, and provides a platform on which third parties can create applications that use the system as well. Open Graph Stories consist of 4 pieces of information: the person, the action, the object, and the domain. Spotify, for example, would say “Mark Zuckerberg listened to Rihanna on Spotify.” Instead of requiring that individual to perform the action of liking the artist’s page, or place it in their profile, Facebook can passively aggregate their stories, and extrapolate the users likes and interests from that. Essentially, these passive actions create a more powerful and accurate concept of interest, with far stronger context than clicking a button.
It may sound creepy, but you’ve all been doing it across hundreds of sites and services for years. It’s not as though you don’t receive utility from contributing to your own online presence. Last.fm has been providing accurate music recommendations to its users for years, as it uses the information gathered from your listening habits and similar people’s listening habits to gauge which new artists you may like. Users of crowdsource services like RunMap.net have been sharing all the best, most scenic runs in their areas. The weakness of these services is that they’re independent platforms. I don’t have a reason to check RunMap unless I’m looking for a run. Facebook, on the other hand, has provided a universal platform on which to create these same services, on top of a base product that is addicting to consumers.
Moreover, the opportunities for Open Graph Stories are unlimited. Even intangible concepts can be broken down into a Person + Action + Object statement. When used in conjunction with aggregation of posts and statuses, Facebook will be able to index tastes and interests that you didn’t even know you have. In fact, if anything, the “like” structure is the biggest weakness in Open Graph, because the context is so loosely related. You don’t know why someone liked something? Was it due to being employed there and their boss demanding they like it? Did they like a book they didn’t read because they enjoyed the movie? Is the food actually tasty at that hole-in-the-wall restaurant, or did they think the name was clever? There’s a difference between anonymous recommendations (Yelp), “likes” (current Facebook), and Open Graph stories that have your best friend eating at the same food truck every day. The recent addition of ratings and intent allow providers to create degrees of interest, as well as declared intent to purchase. The ratings will provide even stronger correlations to their actions, and the you can not get any lower in the sales funnel than declared intent to purchase. Aggregated actions, ratings, and intent to purchase are a far more consistent source of tastes and interest than the soon outdated Like button.
In addition, items that were previously difficult to search for will be accessible. My favorite example is when a friend checks-in with a post saying she is hiking in the hills near San Francisco. This check-in is converted to a story, “Sarah hiked the Tennessee Valley Trail.” A third-party hiking route application shows me the route she traveled. In the future, I can search for “Friends who hiked in Marin County,” see her story, and then view the panoramic photos she uploaded that day. I’ll take my hike, and then send her a thank you while uploading my photos later that day. Facebook is providing a truly useful service that is not currently offered or available in the same form, breadth or accessibility. Graph Search’s biggest opportunity is not the data already in its system, but the fact that as an accessible platform; there is an infinite amount of integrations that would improve social context and accuracy in search. Open Graph will succeed because it provides genuine value to Facebook users, encouraging them to take part in the platform.
Although there is currently no ads product within Graph Search, renewed interest in localized online advertising for small business and ad implementation is a clear next step. Graph Search will put Facebook closer to the end of the sales cycle, yet in a way with which Google cannot compete. For example, my friend posted the story about the hike. An outdoor equipment retailer could use its advertising budget on Google, purchasing ad space for people searching “hiking boots” on Google, but it’s unlikely they’ll beat out online retail giants of Amazon, Zappos, etc. However, if that same budget was spent on Facebook, targeting people who enjoy hiking, live within 25 miles of my store, purchase shoes often, shop offline, have specific hiking brands they prefer, and have a friend who likes my store, the specificity and Referred Intent (i.e., the influence on a user to drive action generated from social context) will lead to significantly higher ROIs.
The retailer is able to target the most likely consumers in keyword areas it couldn’t target before. If there’s someone who has a connection to their store, they’ll be more likely to build a lasting connection with the new customer, much in the same way a warm introduction will always beat a cold approach. Graph Search, and the development of Open Graph will provide higher returns and lower costs for local businesses. Although large companies are currently the bulk of Facebook’s advertising, far more local businesses will be in a position to have future success.
The future of Graph Search is bright because the future of Open Graph is even brighter.