Purpose: This study tests whether consumer-generated electronic word of mouth (e-WOM) has a stronger relationship with local perfume purchase decisions than celebrity endorsement, while controlling for perceived price. Method: The study uses 119 complete survey responses and 19 five-point Likert items. Construct means are analyzed through multiple regression with standardized coefficients and HC3 heteroskedasticity-robust standard errors. A 5,000-replication bootstrap evaluates the difference between the e-WOM and celebrity-endorsement coefficients. Finding: Perceived price is the only significant predictor (β = 0.743, p < 0.001). Celebrity endorsement (β = −0.092, p = 0.353) and e-WOM (β = 0.032, p = 0.794) are not significant after the predictors are considered together. The bootstrapped e-WOM-minus-endorsement contrast is also nonsignificant (95% CI = −0.233 to 0.441). Novelty: Instead of examining promotional variables separately, the study directly compares paid celebrity influence with consumer-generated information for a sensory experience product. The findings challenge the assumption that stronger average e-WOM scores necessarily represent a stronger unique purchasing effect and identify perceived value as the dominant driver.
Copyrights © 2026