Twitter activity predicts viewers’ engagement with TV shows, says study

Nielsen finds tweets can also show which premieres will attract big audiences

Twitter might not be a window into the soul, but it does provide a clear view into what grabs our attention on television, according to research being published this week.

Nielsen, the data company best known for its TV ratings, hooked up about 300 people to brain monitors and followed their reactions as they watched new episodes of eight prime-time TV shows. Researchers then compared those measurements with the volume of messages on Twitter about the same shows when they were broadcast.

The conclusion: The number of tweets correlated closely with viewers’ depth of engagement with what was appearing on the screen at that moment.

“You can use the Twitter activity to predict the engagement of the show,” said Dr Avgusta Shestyuk, director of neuroscience at Nielsen Neuro, the division that conducted the research.


“As the audiences are getting more engaged with the segment, the Twitter activity is getting more intense.”

Nielsen's findings support Twitter's long-standing contention that the real-time, public orientation of its social network makes it a reliable barometer of the public's changing moods and interests. Twitter has about 63 million active users in the United States and 288 million worldwide, although its users tend to be younger and more knowledgeable about computers than the population as a whole.

Special events

Twitter routinely promotes data about the volume of tweets around television special events like the

Academy Awards

, as well as regularly scheduled shows like

The Voice



, as indicators of the programmes’ popularity. It also uses that information to sell ads to brands that want to reach the show’s audience, either while the show is on the air or shortly afterwards.

Now, Nielsen is saying that Twitter chatter is an accurate indicator of the overall audience’s interest in a show, right down to the specific scene.

Other data unconnected to this study also suggests a correlation. The 40th anniversary special edition of US show Saturday Night Live last month, for example, was the highest-rated episode of the show since 2004 and also its most popular on Twitter, according to Nielsen data shared by Twitter. Twitter activity peaked during Kanye West's live debut performance of the song Wolves with Sia and Vic Mensa.

The level of Twitter activity about forthcoming shows before a new TV season begins can even be used to predict which premieres will draw the biggest audiences, according to another study that Nielsen released in January.

(Nielsen did not quantify the impact, but earlier research by the US Council for Research Excellence found only a modest effect compared with traditional promotions such as TV ads.)

Twitter declined to comment on the latest Nielsen research. But the findings, while not surprising, are good news for the company, which is battling with Facebook for audience attention – and the related digital advertising dollars – during the broadcasts of live events like sports and awards shows, as well as regular series.

So far, however, Twitter and Nielsen have avoided answering the most important question posed by marketers and the TV industry: Exactly how much does chatter on Twitter lift the viewership of a particular show? Although Nielsen publishes data on the Twitter activity surrounding a show’s broadcast as a complement to its more familiar TV ratings, it has said little about the relationship between the two. (Even less is known about Facebook’s impact, since most posts on it are private.)

Brain activity

For the most recent study, Nielsen put sensor-filled caps on the heads of nearly 300 people, 21-54 years old, in the

San Francisco

Bay Area, Chicago and Atlanta, and measured brain activity related to emotion, memory and attention.

The company said the group was equally divided by gender and representative of the racial and ethnic make-up of the population in each region. The research participants were all regular watchers of the programme they saw, and many did not have Twitter accounts. While watching the show, the participants were not allowed to tweet, surf the web or otherwise multitask.

The Twitter analysis was based on tweets, retweets and replies posted during East and West Coast broadcasts of eight hour-long shows last fall. Nielsen declined to identify the specific shows studied but said they were from various genres and included both high-and low-rated offerings on cable and broadcast TV.

The researchers found a 79.5 per cent correlation between the level of activity on Twitter and the viewers’ neurological engagement with the show.

Dr Shestyuk said that Nielsen did not examine the audience’s reaction to the ads in this study, although the research suggests that a show’s producers and advertisers could study patterns of Twitter traffic and figure out what resonates most with TV viewers. Other research shows that people tend to remember ads that run during the most engaging shows better.

Testing audience reaction to ads is a big part of Nielsen Neuro's business for brands, and the lab, based in Berkeley, California, plans to conduct a follow-up study to see if Twitter chatter about specific TV ads is a good indicator of how big an impact they have on the people who see them. –(The New York Times 2015)