top of page


Research: Projects

with Puneet Manchanda and Eric Schwartz

Free ad-supported streaming of on-demand content is growing. Platforms that provide this service need to better understand how ad delivery affects the consumption experience – do viewers zap (leave a program incomplete) during immediate ad exposure or subsequent (program) content exposure? Using debiased machine learning with Hausman instruments, we capture the causal effect of four levers of ad delivery on immediate ad zapping and subsequent content zapping behavior. Our results find that on average, an increase in number of pods (ad breaks), length of pods or repetition of ads, results in a larger increase in subsequent content zapping than immediate ad zapping. On the other hand, an increase in spacing till the next pod, results in a larger decrease in immediate ad zapping than subsequent content zapping. We also investigate differences in effects across sub-types of zapping: (a) switching to another episode of same TV show, (b) switching to new TV show or movie and (c) stop watching. We find that the platform can face tradeoffs between preventing ad zapping or content zapping for some ad delivery levers based on whether the platform wants to promote stickiness to the content or to the platform.

Keywords: Streaming Platforms, Ad Delivery, Zapping, Causal Inference, Debiased Machine Learning


with Puneet Manchanda 

Influencer marketing has become a very popular tool to reach customers. Despite the rapid growth in influencer videos, there has been little research on the effectiveness of their constituent features in explaining video engagement. We study YouTube influencers and analyze their unstructured video data across text, audio and images using an “interpretable deep learning” framework that accomplishes both goals of prediction and interpretation. Our prediction-based approach analyzes unstructured data and finds that “what is said” in words (text) is more influential than “how it is said” in imagery (images) or acoustics (audio). Our novel interpretation-based approach is implemented after completion of model prediction by analyzing the same source of unstructured data to measure importance attributed to the video features. We eliminate several spurious relationships in two steps, identifying a subset of relationships which are confirmed using theory. We uncover novel findings that establish distinct associations for measures of shallow and deep engagement based on the dual-system framework of human thinking. Our approach is validated using simulated data, and we discuss the learnings from our findings for influencers and brands.

Keywords: Influencer Marketing, Video Advertising, Social Media, Interpretable Deep Learning, Transfer Learning

bottom of page