Personality does not directly predict brand advocacy in a utilitarian service category: a bounded null, three converging accounts, and the limits of testing cross-brand transfer

Abstract Personality traits are reported to predict brand loyalty and advocacy, and segmentation is sold on that claim. In an analytical subsample of a quota-representative Finnish telecommunications survey ( n = 836), using a general-domain Big Five inventory with no item overlap with the outcome, no trait predicted brand advocacy intention, unconditionally or with emotions controlled. The null is informative rather than inconclusive: the study was powered to detect small effects, and equivalence tests and Bayes factors bound the direct trait–advocacy association below the effect sizes reported in prior work. Discrete brand emotions, measured in the same respondents, explained substantial variance in advocacy. Three accounts predict the null: the Affect-Infusion Model, on which personality modulates rather than generates evaluation; the frame of reference of a general-domain instrument; and the utilitarian structure of the category. They are observationally equivalent in this design, and only the Affect-Infusion Model yields a prediction testable within the study. Exploratory analyses recovered personality × emotion interactions and one brand-level trait coefficient, none surviving correction for multiplicity or robust estimation. Across three competing brands, the standard toolkit for testing cross-brand transfer cannot resolve whether structures differ at realistic sample sizes; pooling is defensible but is not evidence that competing brands are alike.