Having an attractive model shill（充当雇用骗子） for a product only helps influence sales in certain situations, according to a new study in the Journal of Consumer Research. It seems it all depends on the set-up for the advertising¹. "Sometimes attractive endorsers³ increase persuasion⁴, sometimes they decrease persuasion, and sometimes they have no effect at all," write authors Janne van Doorn and Diederik A. Stapel (both Tilburg University, the Netherlands). In four experiments, the authors demonstrated that context is everything when it comes to evaluating the role of the attractive spokesperson.

The authors found that the looks matter most when an attractive person serves as a cue just before an ad with attractive products. "When beauty is used as a cue, the attractiveness of the person is likely to have a relatively⁵ general impact and affect evaluations⁶ of advertised as well as non-advertised products," the authors write.

When an attractive person and the advertised product are explicitly⁷（明确地） linked in the same visual frame, consumers respond in a more specific fashion, and just the evaluations of the advertised products are affected⁸.

Finally, when endorser² attractiveness is used as an argument (for a beauty-related product, for example), the impact depends on the perceived self-malleability of consumers. "Consumers who believe they can improve themselves may see the attractiveness of endorsers as a relevant argument for buying the advertised and non-advertised beauty products and thus evaluate them relatively positively," the authors explain.

When consumers believe their self-image has no room for improvement they will not see an endorser's beauty as a persuasive⁹ argument for buying an advertised product. "If one does not believe in improvement of one's appearance, what is the use of beauty products?"

"The research shows that the effectiveness of using attractive models in advertising or other promotional activities depends on how this attractiveness is used in the design of these activities," the authors conclude.