Best Paper Award for Daniel Bruns, Steffen Prior and Tobias Langner at the AAA 2023 in Denver
Prof Dr Martin Eisend, Prof Dr Tobias Langner and Junior Prof Dr Daniel Bruns (from left to right) Photo: American Academy of Advertising
Marketing Chair researchers Daniel Bruns, Steffen Prior and Tobias Langner were honoured with the Best Conference Paper Award for their paper "Influencer Marketing Effectiveness: Automated Measures of User's Social Media Engagement toward Influencer Posts as Indicators of Attitudinal and Behavioral Outcomes" at this year's American Academy of Advertising (AAA) conference in Denver. The award was presented by Martin Eisend, Vice President of the AAA.
The AAA conference is the most important advertising research conference in the world and takes place annually at different venues.
Influencer Marketing Effectiveness: Automated Measures of User's Social Media Engagement towards Influencer Posts as Indicators of Attitudinal and Behavioural Outcomes - Abstract:
Measuring the success of influencer campaigns is considered to be the most important challenge of influencer marketing, according to recent surveys among practitioners. The present paper provides a comparison of different approaches that measure campaign success based on freely accessible data: engagement rates (like-follower-ratio, LFR, and comment-follower-ratio, CFR) as well as sentiment of user comments analysed using lexicon-based approaches (LIWC and VADER) and a machine learning Naïve Bayes (NB) sentiment classifier. We examine the convergent, discriminant, and predictive validity of these metrics. Linear mixed-effects regressions reveal that only the NB classifier and LIWC converge positively with a measure of attitude towards the post. CFR correlates negatively with post attitude, indicating that lower overall attitudes towards an influencer marketing post yield more user comments. Predictive validity with regard to brand attitude, purchase intention, price premium, and positive word of mouth was found only for the NB classifier and partially for VADER. These findings suggest that engagement rates and lexical approaches provide only limited validity to indicate post attitude or to predict the attitudinal and behavioural outcome variables. The machine learning approach of the NB classifier presented here can serve as a valuable performance indicator and helpful controlling tool for marketers.
