Social Media Talk May Correlate To TV Cancellations

by Wayne Friedman , May 12, 2015

Can social media predict which TV shows might be cancelled?

Networked Insights, an analytics, audience-centric company that focuses on brands, says shows that ended up being "cancelled" had a low 'love' index --- meaning not much in the way of social media conversations where there were high volumes of sentiments/words like "love," "like," "hope" and other emotions.

Research said these eventually cancelled TV shows (or ones that will be) had a low 'love factor,' the "conversation index for emotional attributes," a 50 index. On the flip side, shows that weren’t cancelled had a high 'love' factor index of nearly 120.

Social media conversations that had a high level of 'disapproval' -- shows that were eventually cancelled -- had a 60 index. Those shows that survived had a 40 index. Words contributing to the 'disapproval' factor in social media included "hatred" emotions and sentiment -- and low number of more positive feelings.

Gleaning data from last year’s historical data, Networked Insights says shows likely to get cancelled or to be discontinued this fall season include:

ABC’s "The Whispers," "Fresh Off the Boat" and "American Crime"; Fox’s "The Last Man on Earth"; CW's "The Messengers"; Netflix's "The Unbreakable Kimmy Schmidt"; Cartoon Network’s "Beware the Batman"; ABC’s "Marvel’s Agent Carter," TBS' "Cougar Town"; and ABC’s "Secrets and Lies."

Networked Insights analyzed data from April 2014 to December 2014 for all the social media conversations -- more than 1.5 million social posts -- of the TV shows coming out of last year’s upfront presentations.

audience targeting, social media, television, tv

8 comments about "Social Media Talk May Correlate To TV Cancellations".

Check to receive email when comments are posted.

Ed Papazian from Media Dynamics Inc, May 13, 2015 at 9:54 a.m.
An even better correlation with TV show cancellations is low ratings.
Reply

Henry Blaufox from Dragon360, May 13, 2015 at 10:13 a.m.

Ed Papazian - can the industry use real time social medi sentiment tracking to predict low ratings, or is the rating process and reporting fast enough that it doesn't matter?

Ed Papazian from Media Dynamics Inc, May 13, 2015 at 11:11 a.m.

@Henry, the national TV ratings, not counting delayed tune-in, of course, are available almost immediately. Unlike social media avtivity, which represents only a tiny percentage of a telecast's total audience, the TV ratings measure all of the audience. Also, the social media indicators, however numerous they may be are haevily skewed towards the lightest viewing---18-34-year-old segment of the audience for many shows and, hence, are not representative of their overall appeal. In addition to the rating surveys, the TV networks utilize other indicators---like the ongoing TVQ studies, which ask people if they are familiar with individual series and then get a fix on how well they like the shows. This form of research, while available on a less timely basis than the ratings, has often been shown to have predictive value---if used with care.

Suzanne Sell from Independent, May 13, 2015 at noon

Several years ago, I met with a social media monitoring company that claimed to be able to predict a drop in ratings by the tenor of the online conversation. It's a reasonable assumption that the two might e related—IF the show's audience is of a demo that's socially active. At the time THE GOOD WIFE was very highly rated and produced virtually no social conversation due to its audience demos.

Since then, we've tracked a lot of series in social media. We've found two things--first, that sentiment is very tricky to categorize, and second, that as a series progresses, the conversation becomes more and more positive as those who don't like the show drop out of the conversation AND the audience. So... I'm with you, Ed!

Dan Neely from Networked Insights, May 13, 2015 at 1:25 p.m.

Hopefully we have moved past demo based audiences at this point... Mom,s Dad's, millennials, New parents, empty nesters, Athletic Females those are the opportunity audiences brands care about and want to market to and they no longer want to settle for proxies. Social Data no longer skews the way in which you describe it. As long as you consider all commentary or content created by a consumer social media you can build an opprotunity audience and understand their entire lives from TV watching to what they eat.

Suzanne... Couldn't agree with you more on the sentiment front.. its a unicorn for unstructured data and one of the reasons we built what we call "THE BRAIN" that understands human emotions across 50+ different emotions... human emotion is not black and white-- positive or negative.... Sentiment as a measure has no real value because most the real emotion is disguised in the Neutral... TV viewers express love, happiness, joy, disgust, confustion, intent.... and they do not do it by using these words in their posts... so we had to build a brain that understands human emotion linguistically.

Ed... knowing i am not going to change your POV on the topic-- I do offer the following... Surveys are by nature highly biased... the people that take them, the leading nature of the questions... surely Heisenberg's uncertaintly principle has some bearing on surveys or panels versus that of how data is contributed in social.

The traditional methods of acquiring data have changed but the tools have not... surveys and panels used to be door to door then to telephone and then to the internet. The tools have not changed, we have merely changed the mechanisms solely because of cost to scale. What has changed is the massive amounts of data available, such that any question that could possibly be asked has already been answered by consumers. Lastly we have to consider the changed consumption habbits of TV watchers and the pressuce that places on the current approach. Not being able to include delayed tune-in in real time becomes the weakest link in the current tools and will lead to a decay in accuaracy as consumption patterns continue to change. Static systems cannot evolve to the new consumer habbits fast enough.

Judit Nagy from FOX, May 13, 2015 at 2:37 p.m.

Social TV comments, which occur mostly while live TV shows are being viewed, are a 'natural extension' of the viewing experience. Someone ‘disrupting’ their live TV viewing experience by jumping on social media platforms to express their feeling, thoughts, love, and ideas about just what's going on right front of them is a big deal. Marketers pay Billions of dollars every day to get some kind of an action, click or WOM generated around their brands, day in and day out.

So, let's give some credit where credit due to the millions of live/on-air comments and the billions of impressions those live comments generate moment by moment, 24/7 on various social platforms. These 'TV shock-waves maybe coming from the 18-34 group, or I would even say even from the 13-34 group, which is the future generation of TV viewers or say 'long form video content viewers'. While they do not represent all of TV viewers, they are serious indicators, 'building blocks' for the entertainment/TV industry.

By nature, these comments are 'real-time' and cannot be immediately applied to in the moment storytelling, but it can be taken into consideration with programming, marketing, and generate genuine engagement around live TV shows.

Just to stay on topic, emotionally filled Social TV comments, can point to engagement with a particular show, and at the end of all, what these comments can do is either tell us that current viewers coming back to Live TV tune in, and also these comments act as 'recommendation' agents to others, creating curiosity and increasing Live or non-linear viewing of a show.

Douglas Ferguson from College of Charleston, May 13, 2015 at 2:37 p.m.

Dan... Surveys are biased, but ratings are not surveys. They measure viewing behavior, not attitudes. Yes, they're flawed, but I have to side with Ed, even though I love social media. Ratings are the best indicator of success.

Ed Papazian from Media Dynamics Inc, May 13, 2015 at 3:18 p.m.

@Dan, the available data continues to show a very strong demographic signature for social media that skews it sharply towards the younger end of the scale as well as other demos. It is not, by any means, an across the board activity. In that sense, it presents a biased view of viewer reactions since the findings are, in most cases, not projectable to a program's total viewing constituency.

That said, were I in charge of TV programming at a network or cable channel, and concerned about cancelling or renewing a series, I would look at anything that might be relevant---the ratings, certainly; the TVQ scores, of course; plus the letters and emails that often pour in; what the critics---those who seem qualified---have to say; any testing results I may have----and, without doubt, I would want a report on the social "buzz" about the show---- how many are commenting, who are they by demos, what they are saying, how this trending, etc. I would also examine the performance of the show's leadin on my channel in the same manner as well as how the major competing entries are doing. Finally, I would look to all of the sources to tell me whether the genre my show happens to be in is still popular---to see if some broad based shift in public attitudes is taking place regarding this type of program content and, if the data were available, I would track the response to my show's stars in the same manner. How do they fare when they do guest appearances, what's the "buzz" about them, etc. With all of this to guide me, I would probably make the right decision; limited to only one indicator, I'd have to stick with the ratings.