Brain Scans Help Predict Ad Success

February 12, 2015
Behavioral Science, Marketing

What kind of ads actually make you buy things? No, wait, don't tell me. Just let me scan your brain.

A team including Duke University Fuqua School of Business Professor Bryan Bollinger has identified a measure of brain activity that can help predict how effectively a TV ad will yield increased sales.

The group used functional MRI brain scans to reveal that activity in the ventral striatum, a reward center deep inside the brain, outperformed other neurological measures in predicting ads that were successful in the marketplace.

"It's the first study that maps actual performance in the marketplace with multiple neurophysiological measures of different ads," Bollinger said. "The concept was very basic. It was to answer the question, do these neurophysiological measures all help explain advertising effectiveness?"

Bryan Bollinger

Bollinger was part of a research group that included professors from Temple UniversityNew York University and the University of California, Los Angeles. Their findings, "Predicting Advertising Success Beyond Traditional Measures: New Insights from Neurophysiological Methods and Market Response Modeling," are published in the Journal of Marketing Research.

The researchers showed 277 people some combination of 37 commercials lasting 30 seconds each, for 15 brands from six companies. The non-profit Advertising Research Foundation, which sponsored the study, provided ads for products ranging from financial services to car rentals to laundry detergent. All of the commercials had previously run on TV, and the foundation also supplied researchers with sales data.

In the lab, the reaction of participants were measured in several ways: By the traditional pencil-and-paper method of asking which ads people liked; with a brand association test; eye-tracking; biometric responses such as heart rate and sweat production; and two types of brain scan, an electroencephalogram (commonly known as an EEG, which records electrical activity from the brain's surface) and functional MRI, which monitors changes in blood flow correlated with activity throughout the brain.

"The main question we focused on was how much additional variation in ad effectiveness can be explained by these neurophysiological measures," Bollinger said. "If you were to pick one method, which one explains the most effectiveness, above and beyond paper and pencil measures?"

The group discovered the more activity found in the ventral striatum, the more successful the ad was in the marketplace, based on data the companies had gathered on how well the ads boosted sales.

"Asking people how much they liked the ad best explained the ad's effectiveness," Bollinger said. "But the fMRI measures collectively explained the most beyond that. The ventral striatum signal was the specific variable that turned out to significantly predict increased advertising effectiveness by itself."

The study suggests an untapped area for marketers who use traditional questionnaires to predict the effectiveness of an ad before it runs. Bollinger said that while fMRI scans are expensive, marketing companies may want to think harder about using them.

"We provide preliminary evidence that some of these measures have predictive ability in explaining ad successes, in particular fMRI," Bollinger said. "But we're not saying these other measures are not useful in explaining ad effectiveness. It could just be we don't have enough data to identify the effectiveness of these other measures, or that some of these techniques are useful in explaining other behavior beyond the simple sales response. The untapped potential of these different methods in better understanding consumer behavior is incredible — we have barely scratched the surface."

This story may not be republished without permission from Duke University's Fuqua School of Business. Please contact media-relations@fuqua.duke.edu for additional information.

Contact Info

For more information contact our media relations team at media-relations@fuqua.duke.edu.