Facebook announced it was opening anonymized conversations for data analysis in March of 2015 and social listening platforms have since been scrambling to offer the feature. They’re forced to go through Facebook’s partner — DataSift — for access so seeing conversations from the world’s largest social network is typically available for an additional cost.
But the only thing listening platforms could offer previously was data scraped from publicly available groups and brand pages. The reality of that situation was that the lion’s share of social media conversations were never seen nor analyzed by your listening platform. In fact, the difference in volume of posts by people vs. brand pages I found while exploring conversations on Facebook about craft beer recently was startling.
With 233,000 conversations coming from Facebook users and only 20,000 coming from Facebook pages, the indication is standard listening platforms are missing 90 percent of the conversations on Facebook. In the average topic analysis, I’m guessing that would account for 40-60 percent of all conversations online. Social listening platforms simply aren’t delivering the majority of social conversations, but because Facebook wouldn’t let them, until last year.
Jay Baer’s new book Hug Your Haters reinforces that assertion. According to his research — a representative sample study conducted by Edison Research for the book — 70 percent of social media complaints occur on Facebook (compared to only 17 percent on Twitter). Yet when you look at the sources in most listening platforms you seldom see more than a sliver of conversations coming from the world’s biggest social network.
So the fact that listening platforms like NetBase, NUVI, Sysomos, Synthesio and others now offer Facebook Topic Data is fantastic. But there are two big asterisks.
First, instead of pulling Facebook Topic Data in to just fill out the data coming from Facebook as a source (so it’s considered along side data from Twitter, News Sites, Blogs, etc.) most platforms are presenting it within its own interface or feature set. So you can’t look at the total conversation. You look at what you’ve always gotten, then you have to go somewhere else to analyze Facebook.
The other is that because these platforms have to pay for the data they pull from DataSift/Facebook, Facebook Topic Data is generally presented as an add-on cost, meaning your potentially already expensive listening and research platform is going to be more so if you want Facebook, too.
On one hand, I understand. DataSift charges the software companies and it would be fiscally irresponsible to not pass that cost on to the end user. On the other hand, not presenting the totality of conversations together is, in a word, dumb. (Though I suppose there could be some data structure inconsistencies that cause this, but it shouldn’t be anything some good engineers couldn’t figure out.)
Perhaps my perspective is looking a gift horse in the mouth. We didn’t have access to Facebook data at all until March of last year. Am I wrong to complain that the listening platforms aren’t delivering it ideally? You be the judge I suppose.
But if you are a brand wishing to use social listening for mining consumer insights or market research, you’re going to have to pony up and work around the clunkiness of Facebook Topic Data for now.
Have you found a platform that marries Facebook Topic Data with other sources well? I’d love to hear about it. The comments, as always, are yours.
NOTE: Image pulled from NUVI. Test account provided by Firebelly Marketing.