More Bang for Your Buck: Effective Kol Marketing Campaign in Emerging Short-Video Markets
36 Pages Posted: 25 Aug 2020
Date Written: July 19, 2020
With the rising of social media, firms seek to contract with key opinion leaders (KOLs) who have strong influence on their social media followers, to promote products of the firms by making videos. This paper studies the problem of how to select KOLs and schedule their advertising campaigns in order to maximize advertising effectiveness. This research augments the existing literature by investigating the distinctive KOL marketing in social media context. For practitioners, this paper provides firms with insights and practical tools for KOL selection and scheduling with varying levels of advertising budget. Based on a dataset from a popular short-video platform in Brazil, this study proposes what we call a multinomial logit model with decaying utility (MNL-DU) to model viewers' choices of consuming the promotion videos. Using this model, we develop a heuristic to solve the selection and scheduling of KOLs' advertising campaigns. Theoretically, we prove that the performance of our proposed heuristic is near-optimal and analyze its asymptotic behavior. Empirically, we estimate this structural model with real data, and further obtain several important empirical results derived from the heuristic, which provide managerial insights for practitioners. According to our empirical results, popular KOLs (who have a larger number of video views but a higher cost to contract with) have significant influence on the viewers. However, we find that only contracting with them will lead to significant loss in revenue even with relatively large budget. Our results suggest that a number of alternative factors are crucial in selecting who to contract with, such as whether the KOL has unique styles, how positive the KOL is rated, and how active the KOL is. These factors also affect the scheduling of the selected KOLs.
Keywords: Social Media; Key Opinion Leaders; Advertising Campaigns; Multi-Nomial Logit Model; Selection; Scheduling
Suggested Citation: Suggested Citation