Every creative review ends the same way: someone says an ad 'feels right', everyone nods, and the campaign goes live. Emotion is treated as the one part of creative effectiveness too subjective to measure – a matter of taste, settled by instinct.

That assumption is expensive. The evidence shows emotion is one of the most commercially powerful levers in advertising, and most of the tools built for measuring emotional response in it don't hold up.

Why emotional response drives more than marketers assume

The commercial case is not in dispute. Analysing 996 IPA Databank campaigns from 1980 to 2010, Les Binet and Peter Field found that campaigns built around emotional appeals were almost twice as likely to deliver top-box profit growth over the long term as campaigns built around rational, information-led messaging. Emotion doesn't just make an ad more pleasant to watch – it's a driver of business results in its own right, and one that most effectiveness conversations still treat as a nice-to-have layered on top of the “real” metrics.

Read: The problem with A/B testing on platforms like Meta and Google 

Why facial coding doesn't measure it reliably

The problem is what happens when a brand tries to actually measure emotional response. The most established commercial technique is facial coding – tracking a viewer's facial muscles while they watch an ad and inferring emotion from the expressions detected. It sounds rigorous. NielsenIQ's own research into the method found virtually no correlation (r = 0.07) between the facial expressions a viewer displays and the emotional state measured directly from their brain. Correlations with actual ad-driven sales were similarly weak – 0.18 for positive expressions, effectively zero for negative ones.

Part of the issue is that people simply don't pull many expressions worth measuring: most video ads generate a reliable, codable facial expression for only a second or two across a 30-second spot. Microsoft, Google and the UK's Information Commissioner's Office have each raised similar concerns about the technique's scientific validity.

That leaves most brands with an uncomfortable choice: trust the room's gut feeling that an ad “feels right,” with no way to interrogate why or defend the call later, or invest in a measurement method whose own published data struggles to connect facial expressions to either the emotions people report feeling or the sales those emotions are meant to predict.

A more reliable way to measure emotional response in advertising

Neither option is good enough for a lever this powerful. The way out isn't a better camera pointed at a better focus group – it's measuring emotional response the way cognitive science actually understands it: as an automatic, largely unconscious evaluation that happens before a viewer has any expression to display or any feeling they could describe in a survey. Decades of research into evaluative conditioning show that pairing a brand with positive stimuli measurably shifts how people feel about the brand itself, independent of what their face does in the moment. That's a fundamentally different signal to chase than a raised eyebrow on camera – and a far more scalable one, since it doesn't depend on running every asset through a lab.

Creative teams don't lack instinct for what feels right. What they lack is a way to check that instinct against anything more reliable than the next person's opinion in the room – or a measurement method that, on its own evidence, doesn't measure what it claims to.