A General Method for Estimating Asynchronous Dynamic Models: A Novel Study of 100 Ad Creatives

Edlira Shehu, Daniel Zantedeschi, and Prasad A. Naik, 2019, 19-105-01

To understand how video ad content drives likeability, some market research companies collect moment-to-moment reactions of consumers towards ads. These studies are inherently cross-sectional as they rely on the variation across multiple ads to identify the average effects. They do not offer diagnostic information on specific scenes of a single video ad or which of the ad’s characteristics are associated with the low liking.

Extant approaches cannot yield such diagnostic information because scenes and liking evolve at unequal frequencies. The standard time series analysis requires observations to arrive at the same frequency. In contrast, to conduct a scene-by-scene analysis requires an approach that tackles asynchronous time series.

Edlira Shehu, Daniel Zantedeschi, and Prasad Naik develop a general method to analyze multiple asynchronous time series, which accommodates multiple dependent variables and multiple regressors with either same or unequal frequencies. They furnish proof-of-concept for the proposed method. Using Google’s 30-seconds ad and Apple’s 60-seconds ads, they illustrate how it estimates the ad content effects of an individual advertisement uncontaminated by the presence of other ads in the estimation sample. The results identify not only the specific scenes in the ad to be edited, but also the significant ad characteristics (e.g., entertainment or irritation) for re-visualization.

They apply the proposed method to 100 video ads and conduct a meta-analysis to discover findings that generalize across five different industry sectors.


The empirical results show that the dramatic impact of 60-second ads dwells in the persistence of the flow of liking. Furthermore, the heterogeneity in content effects relates to the narrative elements of plot structures. Indeed, the commonly used plot structure “stick-to-one theme” may not be the only way to build and carryover liking for ads.

Regarding content dimensions, the impact of entertainment, relevance, or warmth on liking can be either positive or negative for a given creative. In fact, the results suggest that warm, stimulating or familiar ads can be hazardous in building liking.

Marketing managers can use the proposed method to estimate ad specific magnitude, direction, and significance of the content effects without being contaminated, as in the extant approaches, by the presence of other ads in the estimation sample.

Edlira Shehu is Associate Professor of Marketing, Department of Marketing, Copenhagen Business School. Daniel Zantedeschi is Assistant Professor of Marketing, Fisher College of Business and Translational Data Analytics, The Ohio State University. Prasad A. Naik is Professor of Marketing, Graduate School of Management, University of California Davis.

The authors thank seminar participants at the McGill University, the Ohio State University, 14th Marketing Dynamics Conference, and 39th ISMS Marketing Science Conference for their helpful suggestions. They gratefully acknowledge the funding and feedback received from the Marketing Science Institute. They also thank the company MetrixLab for providing the data used in this project. The last author acknowledges the financial support received from the UC Davis travel and small research grants program across 2015-18.



  • Corporate: FREE
  • Academic: FREE
  • Subscribers: FREE
  • Public: $18.00



Employees of MSI Member Companies enjoy the benefits of complete online access to content, member conferences and networking with the MSI community.



Qualified academics benefit from a relationship with MSI through access to, conferences and research opportunities.



The public is invited to enjoy partial access to content, a free e-newsletter, selected reports and more.




Learn more about becoming a member of the institute

Read More

Stay Informed

The MSI Mailing List

Subscribe to our email list to stay informed about upcoming events, news, etc.

Social Media