GraphScience launches Facebook ad optimization platform

New service tries to compete with big guns to turn Facebook fans and likes into customers

You've likely not heard of GraphScience. That's because the service has only now officially launched Tuesday on Facebook, even though the Palo Alto, Calif-based start-up has been around since 2010.

From the sounds of it, however, staying relatively quiet on the media front hasn't hurt its ability to make real revenue.

"We have dedicated Facebook budgets ranging from $3 million to $20 million this year," said Raymond Rouf, founder and CEO of GraphScience, in an interview with me. 

So what is GraphScience? "We're a social data company doing Facebook advertising," said Rouf. "We're understanding all the science behind the social graph."

Basically, a customer - mainly a retailer, such as Hautelook  and consumer packaged goods companies - will work with GraphScience in order to buy Facebook fans or likes. Right now, around 30 clients pay GraphScience to manage between $50,000 and $200,000 a month in marketing spend on Facebook to drive customers to a sale. GraphScience gets an undisclosed fee on that spend and, among other things, will analyze the customer's likes to see which demographic and interest group to target. 

"We'll find users they [our customers] didn't know they had," said Rouf. For example, in one case, GraphScience saw a strong correlation between people who liked the Bible and people who liked a client's products. By slicing and dicing the social graph to see clusters of users that were more inclined to like a certain product, GraphScience was able to make that connected. Eventually, they went on to target people who like Holy books and people who liked different sects of Christianity. "We're getting granular in terms," said Rouf. 

In many ways, bidding on these different interest groups is similar to bidding on keyword strings, something SEM (search engine marketing) service providers have been doing for years. 

Now many of the big guns, like Efficient Frontier, Kenshoo, and Marin Software are taking their SEM skills and eyeing Facebook. At the same time, emerging start-ups, like Trada, which has raised more than $15 million in financing, are also eyeing Facebook to be the place to help analyze the social graph. 

To that end, what's GraphScience's competitive advantage?

"We're finding more inefficient targets," said Rouf. "We're finding which interest data isn't being bid on and we're understanding the correlation between the interest data and the product."

GraphScience points to high click-through-rates (CTRs) as one of the reasons they've been able to produce rich returns for its clients. Facebook CTRs were .051% in 2010, down from .064% in 2009, according to Webtrends. GraphScience's average CTR runs 50% or higher than that, according to the company.

History of GraphScience

It all started on a Black Friday in 2010. At the time, some companies were paying $4 to acquire a fan on Facebook. Rouf didn't expect those figures to be sustainable. And, he was right. One retailer had spent $3 million on Facebook, only to drop that budget down to half a million in 2011, he said. GraphScience started working with customers to get them fans and likes for 22 cents to 40 cents. At one point, GraphScience worked with one company and delivered an 8x ROI (return on investment). This means, the company spent $10,000 in advertising and saw roughly $80,000 worth of goods sold, said Rouf. 

Today, GraphScience is targeting a 4x ROI for its customers and it's able to deliver that, so claims Rouf.