Focus groups for TV shows and movies have been around for decades. But despite the dollars and hours spent trying to figure out how people will react to media, people don’t necessarily report what they’re actually feeling.
That’s Matt Celuszak’s take anyway.
Celuszack is the founder and CEO of UK-based startup CrowdEmotion, which has created a software program that can analyse facial expressions. Launched only a few months ago, the company is part of BBC’s incubator program, BBC Worldwide Labs. The British broadcaster is already using CrowdEmotion’s proprietary tech to gauge how viewers react to shows like Sherlock.
“The BBC wants to make high quality content and they want to make it stick,” Celuszack told BI Intelligence, “That’s the reason why they chose us. They want to see if viewers will like their content or not.”
CrowdEmotion’s software can track expressions over time, making it possible to measure mood and specific reactions to parts of a show or story. For producers, this means more intel on how specific scenes or dialogue are received. The tech could also potentially help marketers determine how to craft messages in ads and understand if those messages achieve their intended emotional effect.
Data from facial expression measurements give a more accurate way of determining mood and emotion over traditional rank and score studies, according to Celuszak. The company is also toying with voice recognition analysis as an additional metric. Here’s a screenshot of Celuszak in CrowdEmotion’s “realtime feed,” as it tracks his reactions to a piece of content:
Screenshot of CrowdEmotion’s mood data output. “Media planning can use this … [the technology] predicts appreciation of a show, telling a broadcaster where to sell it. That’s our secret sauce,” he writes in an email accompanying the image.
To calculate mood based on expression, CrowdEmotion’s software uses algorithms and the latest in machine learning technology, a type of artificial intelligence in which a system can train itself using data it collects. Emotion recognition software has existed in the field of neuroscience, but this is the first time a product like this has been developed and used outside of a research setting, according to Celuszak.
“This is an example of data science helping, not fighting, creative content makers,” he said.
The software can be used with any device that has a camera. It can even work with smartphones and other portable devices. In the future, CrowdEmotion wants to track audience reactions to any type of content, not just TV. “We are looking to expand into broadcast, music, gaming and publishing.” Other possible applications for the tech industry include using it for border safety patrol to detect criminals.
This story was first reported by BI Intelligence, a research service from Business Insider. Sign up today and receive Digital Media Insider, a daily briefing for digital media executives, every morning in your inbox.