Emotion measurement that can identify which ads sell

We partnered with Mars Inc. and their Marketing Laboratory at the Ehrenberg-Bass Institute for Marketing Science to investigate what drives video advertising success.



Emotions drive behavior, so we wanted to know if emotion measurement could distinguish between ads with high or low sales lift.


We created the largest Sales-linked emotion data set in the world by combining Realeyes emotions data with Mars Inc. single source sales data, and leveraged our SEWA research to create effective predictive models based on that database.


We were able to distinguish between ads that sell and those that don't with 75% accuracy. Successfully identifying emotional behavior signals which together predicted an advert's sales lift across all the product categories and markets tested means emotions can definitely be linked to advertising effectiveness.


Few research methods have proven links to sales. We want to refine our predictive models to achieve a similar level of accuracy across all markets and categories, with the potential in future to optimize creative selection, editing and budget allocation.


Creative Testing

See where to edit your videos and select the ones with the best performance potential.

Media Planning

Refine your media plan by knowing who to target to and how best to distribute your budget.