Traditional TV ratings measure how many people are watching, but a new ratings system developed by Bluefin Labs measures how viewers respond to television. Led by the head of the Cognitive Machines group at MIT Media Lab, Bluefin Labs says it mapped the “TV genome” to come up with a new set of metrics it calls Response Level and Response Share.
Response level is “the number of commenters for any given episode of a show” based on a 10-point exponential scale. Response share is the percentage of a program’s share of social response within a specific daypart, which roughly maps to the traditional Nielsen “share” metric.
For example, in this sample report (click to see larger) Glee scored a 7.7 response level with a 32.5% share of the social conversation around TV shows in its daypart. In another example, here’s the response levels for the May 24th episode of American Idol compared to several big live events:
- 2011 Super Bowl = 9.2 Response Level
- 2011 Academy Awards = 8.8
- 2011 French Open Mens Final = 8.2
- American Idol Finale (May 24th) = 8.0
“With the introduction of Response Level and Response Share, we can provide new insight into audience engagement in a way that has not been possible before,” said Deb Roy, co-founder and CEO of Bluefin Labs. “With these contextualized metrics in hand, media and marketing executives can interpret all responses to any particular show in aggregate.”
Bluefin is also rolling out Bluefin Signals, a web-based analytics product that allows programmers, marketers and buyers to break down the numbers beyond response level and share on over 3,000 distinct shows with 105,000 individual airings. This example above (click for larger) shows “cross-show engagement” for One Tree Hill fans with other TV shows. Every month Bluefin’s technology fingerprints 2 million minutes of TV and correlates that data with 3 billion social media comments — the “TV genome.”
“With the combination of science and large-scale computing, we are solving a problem that has eluded TV for more than 60 years: how to measure actual audience engagement and not just simple media consumption,” Roy said in our story about Bluefin a few weeks ago. Roy and his MIT team developed “deep machine learning algorithms designed to uncover the relationships between spoken language and context,” Bluefin’s website explains.
While there are several services that measure social interaction around TV — for example, Trendrr which powers part of our social TV leaderboard — Bluefin is the first to take a crack at setting an industry standard with similarities to Nielsen’s TV rating and share. Nielsen offers cross-platform reporting spanning TV, web and mobile devices, but it hasn’t gained traction with a social TV standard of its own.
“We are in-market with multiple broadcast and cable networks, agencies and brands looking at this data day-in, day-out, (and) we feel Trendrr.TV is the industry utility,” explains Trendrr’s Mark Ghuneim, responding to today’s Bluefin announcement. “I think it is very important right not not to confuse the marketplace anymore than it needs to be – IE new metrics and new weighting systems.”
While it’s unclear this early if Bluefin will succeed at setting a social TV measurement standard, it certainly has the mathematical star power. Founder Deb Roy made a splash at TED with his project “Birth of a Word” explaining how he videotaped and analyzed 90,000 hours of footage of his son’s life to understand how an infant learns language. Fitting, perhaps, because social TV is just beginning to say its first words.