Predicting Advertising Sales With Biometrics – And 5 Best Practices We Learned From Them

For years, neuromarketeers have been in pursuit of the buy button. This specific neural pattern ought to align perfectly with a rising slope in the sales curve.

Unfortunately, reality is more complex than that. While specific kinds of brain activity are certainly predictive of purchase and preferences (the nucleus accumbens and frontal asymmetry pop up time and again), the ultimate response seems to vary with content and strategy.

High attention can sometimes be optimal – and sometimes ads fare better when attention dips quite low. Ad liking can be a great predictor for a particular subclass of ads – whereas for other types it serves as nothing more than a broken compass. Even the experience of negative emotion and pain could potentially increase advertising effectiveness, given the right circumstances.

Steven Bellman and his colleagues argue that neuromarketing researchers shouldn’t measure each and every ad against an overall neural benchmark of what makes the ‘perfect ad’. Instead, advertisers need to have a better a priori understand of what makes their ads effective and then research whether their ad succeeds on these KPI’s. One size doesn’t fit all.

Let’s take humorous ads, one of the most common advertising appeals nowadays. A funny ad may work well for chocolate. Not so much for banking. One of the sales predicting KPI’s within this ad category could be: do people actually experience the ad as funny?  If this is the case, a biometric that measures the experience of humor (for instance, through facial coding) becomes a predictor for ad effectiveness within this specific category (but not for all ad categories per se).

Interestingly, a recent field study investigated this very question.

Does facial coding predict sales?

A team of researchers from the Ehrenberg-Bass Institute for Marketing Science probably had little sleep last year. They measured the effectiveness of more than 100 ads by using both biometric and purchase data from over 1,000 respondents.

The researchers did something that has rarely been done in neuromarketing studies: using single source data to predict sales. This means the researchers didn’t just look at aggregate sales numbers from market data, but instead monitored each and every respondent individually – from television to checkout counter. This approach establishes actual causal effects.

All respondents’ reactions to the ads were measured alongside two biometrics. First, facial coding software registered people’s second to second smiling response. Second, heart rate monitoring measured the time interval between beats, which is a measure of attention (indeed, our heart slows down a bit when we focus our attention).

Humorous ads were generally more successful (61% success rate) than non-humorous ads (39% percent success rate). The predictive power of the biometric data was quite remarkable in this regard. The researchers were able to separate success from failure for 69% of the ads. If at least two out of the following 5 rules hold, the ad will likely be successful:

• 1- The ad uses humor
• 2- The ad is actually funny (20 percent of the audience for the ad was smiling for 50 percent or more of the ad’s duration)
• 3- The slope of smiling is positive (which means the number of smiling people increases along with the commercial’s duration)
• 4- During the peaks and final moments, more than 25% of the people should be smiling
• 5- The ad revolves around one of the following humorous themes: imitation, impersonation, eccentricity, sexual allusion, repetition, or grotesque appearance

How to quickly identify poor performing ads before airing

The most pressing question for advertisers is how to weed out poor performing ads before wasting money by airing them. How do you identify the real stinkers?

The researchers discovered the interval between heartbeats – a measure of attention – to be the best predictor in this regard.

The ultimate factor separating weak ads from all the others is that they evoke lower levels of attention. Interestingly, attention seems to be a hygiene factor. Attention needs to be in place, but it doesn’t automatically elevate an ad to greatness either. This is evidenced by the fact that there’s no difference in attention between fair, good and excellent ads. Only weak ads have weak attention.

In particular, low attention during the opening seconds of the commercial is detrimental. So, how to grab attention? The researchers found that the best attention-grabbing snippets contained humor, included a twist or change in plot, or showed close-ups of characters’ faces.

No golden formula

So… are you about to jump on the humor train for your next ad? Hopefully not. At least not without giving it some proper thought.

Although it seems to work wonders for most ads, that’s not the main point this research wants to underline. Advertisers have many techniques at their disposal, and they can turn to biometrics to measure whether these techniques work as intended. But which specific tool in the box is most suitable for your ad is entirely dependent upon the type of product and the corresponding mindset and motivations of the viewer.

In this study, humor worked extraordinarily well, possibly because it contained mainly ads for low involvement consumer products. Had the study used other ads – let’s say banking and insurance services – humor wouldn’t be nearly as effective, making the smile response irrelevant to research.

Effectiveness is highly dependent on context. There are no golden formulas. No hypnotic tricks. No buy buttons.