top of page
  • Writer's picturePeter Hammer

What AI tells us is the key to higher social video engagement

Updated: Jun 10, 2020

There’s lots of excitement around artificial intelligence. And we are seeing lots of new applications where machine learning and computer vision technology can be used to enhance our knowledge, uncover new insights and enable us to make scientific discoveries that couldn’t be achieved with humans alone. But how can this be applied to marketing – specifically video marketing?

Last year, we conducted a study that used computer vision and machine learning to code up 497 social videos for Australian brands. We then looked at the different creative devices used in each clip, and how this impacted engagement.

Overall, we found that those social video clips that used rare creative devices (those that did not commonly appear in all the videos we analyzed) achieved an engagement rate 15% higher on average, as compared to those that didn’t use rare creative. It suggests that if brands could identify rare creative devices for their category, it might help them to cut through the clutter and drive more impact.

Social Video Dataset: Engagement Index by Presence of Rare Creative Device Categories

We have only just scratched the surface, but we believe this technology will help to drive better outcomes for marketers and advertisers in terms of video and social media impact.

To find out more about the study – or download a lite version of the report – please visit :



bottom of page