Video consumption decreases when users are presented with long, skippable ads, as per research conducted by VDO.AI.
To study the multi-objective personalisation of the length and skippability of video advertisements, the research was conducted by Omid Rafieian, Assistant Professor at Cornell University - Cornell Tech NYC, Anuj Kapoor, professor at IIM Ahmedabad, along with Amitt Sharma, Founder, and CEO at VDO.AI, under the endorsement of Z1 Media.
According to the research, while video consumption decreased when users were presented with long, skippable ads, ad consumption increased - posing a challenge for platforms seeking to optimise both outcomes.
To address this issue, VDO.AI developed multi-objective personalisation algorithms that utilise individual-level substitution patterns to optimise both ad and video consumption.
The results of the study show that multi-objective personalised policies can significantly improve both ad and video consumption outcomes over single-objective policies.
As per VDO.AI, the algorithm was able to increase ad consumption by 61% at the expense of only a 4% decrease in video consumption. Similarly, compared to the single-objective policy optimised for ad consumption, there is a multi-objective policy that increases video consumption by 47% while decreasing ad consumption by just 13%.
"We are excited to have partnered with VDO.AI on this study and to have contributed to the development of multi-objective personalisation algorithms that can significantly improve both ad and video consumption outcomes over single-objective policies," said Omid Rafieian, Assistant Professor at Cornell University - Cornell Tech NYC.
Anuj Kapoor, Professor at the Indian Institute of Management Ahmedabad, added, "The study's findings have important practical implications for platform decision-making in real-time, and we are excited to have contributed to this important research."
Founder and CEO at VDO.AI, Amitt Sharma said, "We are thrilled to have collaborated with Cornell University and the Indian Institute of Management Ahmedabad on this groundbreaking study. Our multi-objective personalisation algorithms will revolutionise the way digital video advertising is approached, and we are excited to see how they will be implemented in real-time."
The study was conducted using field experiments and utilised machine learning and causal inference techniques to analyse the data.