Facebook’s Graph Search was announced about two weeks ago and so far reactions from the marketing community have been fairly mixed. The new and improved search tool, currently in open beta in the US, lets Facebook users conduct more detailed searches of their connections based on places, photos, likes, locations, interests and updates.
With demographic and interest data from over one billion people, Facebook’s Graph Search should be a marketer’s dream. It has the potential to be a powerful tool for marketers to learn about their target audience and fans, and is great for finding new customers. There is, however, one major problem with Facebook’s Graph Search – “dirty data”.
By now you’ve probably seen the Tumblr blog featuring a number of real Facebook graph searches. A search for “married men that like prostitutes” brings up a long list of – perhaps soon to be divorced – men. The married men revealed by this search were presumably trying to be humorous amongst their friends in a social context, which Facebook is a fine platform for, but what this example highlights is that Facebook’s “brilliant” social graph is not an interest graph. It’s more one of non-contextual relationships and keyword frequencies. The problem with some of the data Facebook relies on to power Graph is that it’s dirty data in the sense that it’s often misleading, and in more ways than one: it can be out of context, factually incorrect or sometimes well out of date.
Since the birth of the ‘Like’, numerous companies like Wildfire have sprung up, enabling brands to set up sweepstakes through Facebook – the preferred entry mechanism being to like a brands page, and herein lies a potential problem. I recently entered a sweepstake for an all expenses paid trip to Italy from a well known low budget pizza chain. I’m not a fan of this brand at all and would never recommend it to a friend, but I liked the company’s Facebook page because doing so allowed me to enter a competition to win a prize that I really wanted. The search results from Graph Search however would paint a different picture. A friend searching for “pizza restaurants my friends like” is likely to bring up a recommendation from me for the aforementioned pizza chain, despite the fact that I don’t really like it. Facebook’s Graph Search just doesn’t understand the context of how a ‘like’ was generated. This is an issue that marketers will have to deal with should they decide to utilise Facebook’s Search Graph – there’s no way for them to distinguish between the consumers genuinely interested in buying a brand, product or service and those who aren’t.
So why is Facebook pushing graph search? Graph search is Facebook’s latest attempt to increase the value of a ‘like’ and encourage more interaction on the social network. The more likes and check-ins (if relevant) that you can generate, the higher you’re likely to appear in Graph Searches and the more recommendations – be those true or false – you will feature in.
If user engagement with Graph Search is high and it becomes as robust as Facebook hopes, there are some big implications for brands. While most marketers recognise the importance of having a social media presence, it’s often only used as an additional avenue to push out announcements. Marketers would have to interact more with their fans and fight even harder to get likes to influence search recommendations. The need to regularly publish high quality and relevant content to Facebook pages would become more important than ever before, so as to attract likes and shares to ensure they appear in all the relevant searches. Social care will also become more important, not only because brands are likely to become more discoverable, if Graph search follows the Edge Rank algorithm currently used to govern the news feed, then comments from fans will also play an important role in deciding content rank.
Should marketers be concerned and start preparing for the eventual rollout of Facebook Graph Search? Combine the current issue of users getting fed up of being bombarded with irrelevant content from brands and the glaring privacy concerns Graph Search highlights, and it’s unlikely that we will see consumers jumping to share more information about the things that interest them, which is precisely what Facebook requires to make graph search really work.
With click-through rates on Facebook being half of typical online industry averages (Facebook S-1, 2011), Facebook desperately needs to prove brand awareness value in an impression or a like and Graph Search is their latest attempt at this. However, a true interest graph, one that links you to the things you are genuinely interested in, is where the true value lies for brands and marketers. Like Facebook, consumers’ volunteering their opinions is important, but it needs to be in the right context.
Henry Lawson, CEO of nFluence.