People went “crazy” over the season seven premiere of HBO’s Game of Thrones.
Or so they say.
“Crazy” accounted for 8.9 percent of all emotional reactions (ERs) during the first episode, and “crazy” ERs spiked during singer Ed Sheeran’s cameo, driving 4.8 percent of all ERs.
What are ERs, and why does any of this matter? Well, if you were the producer of Game of Thrones, you looked at this data to evaluate how the show resonated and determined what to promote across social media over the following days.
The power of emotions, as a way to better connect with audiences and create fans, is the big bet of Canvs, a New York-based tech company that works with 90 percent of the TV networks in America—including HBO.
Just recently, Canvs added Instagram to its platform, which analyzes reactions and posts on Facebook, Twitter, and YouTube in real-time for particular TV shows and online video series, tracking and visualizing the conversation at every second:
CEO and founder Jared Feldman said he sees his platform Canvs as a new kind of focus group.
“We believe emotion is Web 4.0. In the age of the internet, there’s never been a standard unit of love. It’s not enough to know that the TV is on, you have this, you have this,” said Feldman, gesturing to a smartphone and laptop.
Canvs isn’t particularly concerned with what screen people are watching on, rather how they’re feeling when they’re watching. Regardless of how shows are consumed, HBO, NBC, and Fox, for example, are using data from Canvs to determine how their viewers are responding.
Canvs is much more detailed than quickly scanning tweets or Facebook posts because the company has spent its history building a database of “millennial” and “modern day” language.
“We don’t speak properly,” Feldman said. “We make up words. We use emoji.”
Feldman said Canvs reports have encouraged producers to either focus on or pull away from a particular character in a show or star of a reality show.
For the season premiere of Game of Thrones, Canvs tracked 172,000 ERs that spread across love, excited, enjoy, and crazy. Love was the top emotional reaction, according to data pulled from Twitter.
Canvs is not alone in the business of listening on social media and providing reports. ListenFirst, Brandwatch, and Crimson Hexagon are just a few tools that also analyze social media posts.
Where Canvs is unique is it does not just provide a scale of sentiment but narrows in on buckets of emotions. YouTube comments—which are seen by some as just another wasteland of the internet— can now be pulled together.
“We don’t do any sentiment,” Feldman said, referencing his competitors. “We’re obsessed with the accuracy of the emotion, and we can group people by age, location, by feeling. We can tell you that people love you tend to be male, and people that hate you tend to be female.”
So far, Canvs has been focused on TV networks. Next up, Feldman said his team is looking to introduce tools for advertisers and brands. In tune with that introduction, Canvs is also building an “impact” score, where it tracks not just emotions overall but determines which ones drive clicks and shares.
“We want to be more prescriptive,” Feldman said. “We could correlate and predict certain business impacts.”