What do people talk about when they talk about _______ on Facebook? Soon marketers will have the answer.
Facebook will start telling marketers the types of people who are talking about specific topics on the social network as well as the kinds of things they're saying, through a partnership with social data firm DataSift.
This so-called topic data can include the demographics of the people discussing a topic, the specific things they're talking about within a given topic including brand names and products -- and even the sentiment of those posts on a topic. Advertisers will be able to filter within these topics by the location and demographic info of the people discussing a topic on Facebook.
Facebook has come up with a few examples of how brands might use this topic data. A hair product manufacturer might like a peek at the demographics of people discussing how humidity affects hair. A retailer might want to know which articles of clothing people mention in their posts. And other brands might be interested in learning whether people are discussing their brands positively or negatively.
With this data in hand, marketers may be able to familiarize themselves with new customer groups, suss out which products they might want to stock up on or highlight in an upcoming campaign or do some recon on how people view their brands.
Facebook will only be pulling topic data from public or private posts people shared with someone else on Facebook. Conversations conducted within Facebook Messenger will not be included. Comments made on Facebook will only be included as the total number of comments on a given topic, but the content of individual comments will effectively be a black box.
To make this possible, Facebook is opening up its aggregated and anonymized data to DataSift, which will organize it for social analytics companies to be able to ingest it into their marketing analytics dashboards that advertisers use. DataSift will make Facebook's topic data available first through a limited number of social-analytics firms in the U.S. and the U.K. Facebook declined to name any initial partners, and it's unclear when it will open up that information more broadly.
Advertisers won't be able to retroactively search for Facebook's topic data through their social-analytics firms. That is, a brand wouldn't be able to submit a query like "hair and humidity" and immediately see results collected from the past day, week, year or ever. Instead once the brand submits the queery, Facebook and DataSift will start collecting and cataloging eligible Facebook posts on the topic. A query would need topic data from at least 100 unique users in order to return any results.
Facebook won't be exposing any of this topic data through the insights tools it provides marketers, so for a brand to get their hands on it, it will have to go through a company pulling data from DataSift.
Advertisers won't be able to access the raw topic data, such as individual Facebook posts on a given topic. And they will not be able to directly use the topic data for ad targeting, though they may be able to incorporate the aggregated insights when creating campaigns. For example, the hair product manufacturer wouldn't be able to target ads on Facebook only to the individuals who were talking about humidity and hair on Facebook, but that marketer can include the group's demographic breakdown so that the ads are shown to people on Facebook with similar characteristics.