This study examines how social media influencer (SMI) attributes shape endorsement effectiveness across awareness, interest and purchase behavior.
Using 6,633 valid survey responses, this study integrates extreme gradient boosting (XGBoost) with BorutaShap and fuzzy-set qualitative comparative analysis (fsQCA) to identify important predictors and sufficient configurations across funnel stages.
Trustworthiness, informativeness and interactivity consistently emerged as important predictors. Other attributes, including humor, physical attractiveness, opinion leadership, wishful identification, perceived fit and originality, showed stage-specific roles. fsQCA results indicate that awareness and interest can be generated through multiple configurations, whereas purchase behavior requires a more integrated combination of credibility, relational identification, interaction, information value and originality.
Brands should align influencer selection with funnel-stage objectives rather than relying on general popularity or isolated traits.
This study integrates predictive and configurational analyses to advance understanding of influencer marketing dynamics across the marketing funnel.
