- 9th April 2015
- Posted by: Adam Lewis
- Category: Social Analytics
Imagine if everyone in the world kept a daily diary. Sharing the places they had been, things they had liked and not liked and conversations with their friends. Imagine that you could access everyone’s diaries. Then imagine you could analyse all these diaries in real-time and extract insights from it.
That’s not far from what we have with Facebook Topic Data.
It’s possibly the biggest record of human thought ever created.
It’s definitely something that everyone in marketing needs to know about.
What is Facebook Topic Data?
From April 2015, the public API for Facebook data will be turned off. In its place is a much bigger beast. Facebook Topic Data will give access to aggregated and anonymous data via Datasift (the big data company). In other words Facebook is digesting the BILLIONS of posts on their platform to identify keywords and phrases – thus providing insights about conversations on specific topics. It is based on approximately 120 pre-defined categories to include topics such as beverages, cars, politics etc. The new system allows third parties to develop their own queries and pull out the data they want.
Based on the data, you can answer questions like this:
- What share of voice does brand x have vs. brand y?
- What are people saying about my brand and my industry?
- What do the public think of our marketing campaign?
- What are the demographics of the people who are talking about us now?
- When are the peaks in engagement with certain topics?
Why Facebook Topic Data matters to marketers.
1) A massive, unrivalled data-set
The scale of the data is staggering. There are 1.39 billion monthly users on Facebook, resulting in 7 billion likes per day. All other social networks pale into insignificance compared to this. I am guessing that only GCHQ , the NSA and Google can match that for a data-set! All Facebook Topic Data is aggregated, so you never get the original post – thus protecting user privacy. The example that Datasift gives is that instead of receiving raw interaction data such as “I like Coca-Cola!” your result could show that 500 women in Iowa have mentioned Coca-Cola in a positive way. It currently supports seven languages for sentiment analysis.
2) Demographic data you can only dream of
This could be a seminal moment in customer insight. Initially the data is restricted to age range, gender and self-declared location (country, region, state). I imagine this will expand out over time. Just think of the data we (as individual users) are handing over to Facebook: whether we are married, what music we listen to, places we visit, brands we like, our level of physical activity (via things like Strava) and information about our education and work. The list goes on.
Yeah. Social listening tools using Twitter can give you data about gender and location. But. It assumes this based on your name and biography data. What happens if your name is Chris? Is that Christopher or Christina? It assumes that you have identified your country in your biography or that you geotag your tweets (estimated to be less than 5% of users). In contrast Facebook Topic Data is not derived, it is self-declared.
3) A new dawn in hyper-targeted advertising
Imagine if you could target a product offer to people who are talking positively about your product. Imagine if you could serve up a political video to people who were talking negatively about a political party, in a certain area and to a certain demographic of swing voter. At the moment, there is no direct integration between DataSift and Facebook’s advertising platform. But I would be amazed if this doesn’t happen at some point soon.
As Nick Halstead (@nik), CEO of Datasift said at a recent event where he announced their partnership with Facebook: “Its time to say goodbye to big data and say hello to human data.”
What do you think? How could this benefit or hinder you in your role? Please comment or share.