Facial Coding in Advertising Research

An action unit is the movement produced by an individual facial muscle or a small group of muscles, such as the lip-corner puller that pulls the mouth into a smile or the brow lowerer that knits the eyebrows. Ekman and Friesen's Facial Action Coding System catalogs these units so that any expression can be described as a combination of them. Building emotion measurement on action units makes it objective and reproducible: instead of an analyst guessing that someone looks happy, the method records which specific muscles moved and how intensely. Emotion labels are then derived from action-unit configurations, which keeps the higher-level interpretation anchored to verifiable physical movements and allows ambiguous readings to be checked against the underlying units.

Within limits, yes. McDuff and colleagues analyzed thousands of facial responses to ads and showed that ad liking can be predicted from facial expressions, particularly positive ones, with substantial accuracy, and that changes in purchase intent can also be predicted, though less strongly. Importantly, they found that simply making people smile is not enough for purchase intent: positive expressions that occur right after the brand appears are far more effective. So facial coding predicts outcomes best when it captures not just whether people react but when, relative to brand and message moments. It is a meaningful predictor that complements other measures, not a guarantee, and its accuracy depends on video quality and adequate sample sizes.

Not necessarily. Smiles can be polite, social, ironic, or even nervous, and they can occur during an awkward or embarrassing moment rather than a genuinely enjoyable one. Ekman's work distinguishes felt enjoyment, which involves both the lip-corner puller and the cheek raiser around the eyes, from social smiles that may involve only the mouth, so the action-unit detail matters for interpretation. In practice, analysts examine the full configuration of action units, the timing relative to the content, and the context rather than treating any upturned mouth as endorsement. This is why facial coding reports the time course and the underlying units, and why it is often paired with other measures to confirm what an expression really signifies.