Since Graph Search was announced by Facebook, bloggers and digital nerds have been singing its praises and aggressively speculating on what other search engines/sites this new tool will take down (Google watch out! Yelp you’re going down! FourSquare don’t even think about it!). Through the buzz, I have remained cautious, and then I came across an article in Mashable that really piqued my skepticism. Mashable’s Emily Pace used Graph Search beta to find that people who Like Star Wars Like KFC, as well as Coca-Cola and McDonald’s.
In full disclosure, I love Star Wars (the first three films released, not so much the recent ones), but I am not that big a KFC fan. And even if I did like KFC, I probably wouldn’t Like it on Facebook. My realization of this made me aware that there are probably going to be a lot of gaps in the data and opinions people put out there, and therefore the data that Graph Search will have at its disposal. Obviously there are tons of Facebook users, and those users spend tons of time on the site, Like-ing all kinds of things. But that’s just it—you can Like just about anything—a post, a photo, a place, a book, an artist, etc.
People in DC, for example, are much more likely to Like a bunch of non-profits and causes than folks from elsewhere just based on our surroundings. It doesn’t mean we care about certain restaurants or clothing lines any less, but this self-selection and self-reporting of Likes could very possibly skew the data that Graph Search turns up in any given query.
Graph Search focuses on answering questions based on a set audience (defined both by your friends and their friends as well as the privacy settings preventing or allowing sharing), which may or may not be relevant to you in terms of demographics, taste, or location. Between the limited access of data for answering questions and the strong possibility that there are missing variables (I don’t Like everything on Facebook that I like), Graph Search really shouldn’t be lauded as the coolest thing—the rival of Google, the challenger to Yelp.
Yes, Graph Search could eventually become an amazing way to study consumers and their behaviors to better target them through their interests and Likes, and it could even help guide individual users to donate to certain charities, find the best restaurants, or figure out where to stay while on vacation. But in the meantime, the lack of fully fleshed out data on user interests and experiences limits the utility of this new search tool, and this is something that brands and consumers should take into consideration. So perhaps this is the next question to pose: will Facebook users engage in additional “Like-ing” behavior in order to better inform their friends, contacts, and brands about their preferences? Or will users develop “Like-fatigue” rendering their information less than helpful?
I think that there will be enough data points out there that the DC-data skewing effect will be minimal. Perhaps non-profits and causes Pages will see some benefit in placing some Facebook ads, but I don’t see my agency making many.
If you are selling something, you’ll look at any possible correlations to see if you can reach new customers. Are most government agencies looking for new markets? We may want to reach new people, or understand the fans we already have, but I don’t think FB is counting on us to feed their revenue stream directly.
Consider how skewed the data will be now that Facebook is re-using your Likes: http://www.nbcnews.com/technology/technolog/facebook-reuses-your-likes-promote-new-stuff-880942
How many people will stop “liking” things, just because of this?