Machine Learning and Video Advertising

Google is using machine learning to find enough patterns to unlock the ability to figure out how every ad should work with video advertising.

The challenge is that as soon as you make an ad that follows all the rules, then all the rules change.

Data is a tool and like many tools, there’s a lot it can teach us.

Machine learning can help, but it does not replace creativity – it supports it.

The only way to break through with video messaging and creation is to be experimenting all the time.

Currently, they see four primary areas to optimize creative via their video ad experimentations:


  • Tighter framing
  • Faster pacing
  • Large type over video
  • Use of brightness and contrast


  • Build for capturing attention
  • Provide dessert first (AKA, provide a punch in the face)
  • Provide a reason for viewers to pay attention


  • Make creative choices that reflect what you are about and who you are interested in


  • Ad sequencing
  • Best mixture of formats
  • How to employ creative and media assets for maximum impact