Saddle up TV Enthusiasts – It’s the Wild West with Social TV Analytics
October 25, 2011
A recent Mashable post that showcased data from Bluefin Labs reminded me we are still very much in the “wild west” days of social TV, social media analytics, and automated audience intelligence. One needs to look no further than the analysis for the premiere of AMC’s hit show The Walking Dead.
On the surface this looks like an appropriate and accurate analysis. After all, this capability is so new to the marketplace how do people really know how to gauge this data? At Networked Insights we’ve been analyzing social data for almost five years, and we’ve focused specifically on entertainment and television since 2009. So when we look at these kinds of reports certain things just jump out at us on how easy it is to miss the mark with providing real ROI to clients.
Sentiment
Thanks to this being colored green, it is absolutely the first thing that catches my eye. What concerns me though is not the color but the results. Now as we’ve discussed before there are a number of different approaches to semantic analysis, and I’m sure people loved The Walking Dead (I know I did), but when I see 75% of the conversation was positive I was blown away. This is because on average when you analyze conversations across all the social channels (blogs, forums, Facebook, Twitter, etc.), 80% of the posts are typically neutral. This doesn’t mean that they’re not interacting with the content at all; it is just that they are devoid of emotionally charged language. To arrive at such a high number, like 75%, the data that one has to be analyzing has to be of a narrow channel like a single Facebook fan page or following a specific official #hashtag on Twitter. Of course I’m not here to downplay the value that each of those channels contributes to the overall social conversation, I’m just arguing that when you measure data in this way you misrepresent The Walking Dead fans to everyone.
Audience Intelligence
Outside of the numbers featured in this analysis the real heart, or value, for everyone should be in what other shows resonate with The Walking Dead audience. However I find myself struggling here for any real takeaway, because in front of me I’m catered a buffet of cancelled shows. In fact some of these shows are even considered to be huge flops for the networks. Most notably are NBC’s The Cape and Free Agents, whom both had volumes of conversation around how terrible they were. So what is the value then of informing me that the audience I’m interested in hates a bunch of shows? Why don’t you just tell me what their favorite sitcom from the ‘60s was; I think I’d find more value in deconstructing The Walking Dead audience’s nostalgia for Tony Danza than for short-lived flops. What all of this indicates is a lack of curation and a reliance on complete automation that fails at discovering and highlighting actionable insights for brands, agencies, broadcaster, or producers.
The takeaway for content producers, sellers of advertising, or media buyers is that it’s imperative to examine social data in context to your objectives. It’s too easy to publish and misrepresent data, thus cloaking it as social media analysis of TV. The quest for ratings, audience engagement or advertising ROI doesn’t begin from the same position for everyone. Accurate analysis requires sound methodology, leading technology and the right amount of human curation to reveal the value of a television show’s audience in social media.
10.28.11 UPDATE
While updating the links within this blog post we noticed Bluefin Labs has responded to our article and updated their list of shows associated with this social audience. It appears the show filter and sentiment scoring we’re upgraded. We bring this to your attention to explain the difference between the graphic above and the live link.
Comments (1)
[...] important issue: unusually high positive sentiment. I’ve talked about this before in regards to BlueFin Labs and their own unusually high sentiment results. To recap quickly what’s wrong with this, on [...]