The data treasure trove otherwise known as Facebook houses discussions by you and a billion+ other humans on every topic under the sun.
Today, social intelligence firm Synthesio announced a partnership with data provider DataSift to access that Topic Data in order to generate usable insights for brands, political groups, or any other client, as part of a new Audience Analytics service.
A client will be able to query the data and get back insights about discussions on, say, almond milk, Pebble watches, Donald Trump, or any other subject.
The New York City-based Synthesio employs natural language processing and machine learning to detect patterns that can be used for crisis management, campaign planning, brand reputation, or return-on-investment calculations.
VP of product Matthew Zito told me that DataSift, headquartered in San Francisco, is the only intermediary authorized by the social giant to provide access to this Topic Data. Synthesio is only one of three companies that have received approval to use the information.
VB just released Social media management: Tools, tactics … and how to win
$299 on VB Insight, or free with your martech subscription
Facebook, Zito said, wants “to make sure anyone who has access is not a fly-by-night,” and evaluates companies wanting to employ the data on the basis of how it will be used and other factors. Synthesio said it couldn’t provide screenshots of its use of Topic Data for this story because of the Facebook restrictions.
He noted that Facebook topic discussions could be substantially different from similar discussions on other social outlets.
“I might talk on Twitter about how much I love a particular car,” he pointed out, “but could [have a different conversation] when talking with my brother” through Facebook. There’s also a different demographic on Facebook, where users tend to be older than on some other social venues.
While all the data is anonymous and provided only in the aggregate, Zito noted that it does not include any private one-to-one communications on the social platform. It does include such info as status updates, age, gender, information visible to friends or friends-of-friends, shares, and the count and nature of likes and comments.
One might discover, for instance, how many times VentureBeat was mentioned, commented on, or liked, the breakdown of those users by age and location, and the types of words used.
This kind of information can inform research about brand reputation, the depth and range of passion about a new product, which messages are resonating, or the level of purchase intent. This analysis, Zito said, can then be run against those conducted on other sources, like Twitter (although not through DataSift).