Using their findings as a basis for editing commercials, Teixeira and his coauthors show that "pulsing" repeated, brief images of the brand can significantly reduce the likelihood that viewers will zap it, as opposed to showing the brand for long periods of time at the beginning or end of the ad.
"Understanding why viewers stop paying attention to a commercial is obviously very important," Teixeira says. "This is the first research that breaks viewer attention into milliseconds—what viewers see, what causes their eyes to move to particular elements of an ad, the image they were watching when they decided to zap a commercial. We observe at the microsecond level what causes people to get 'ticked off,' stop watching, and press the button."
The eyes have it
Teixeira learned of eye-tracking technology as a Ph.D. student at the University of Michigan. From a participant's perspective, the eye-tracking setup is as unobtrusive as sitting in front of a computer monitor. Eye movements are captured by an optical sensor using infrared light reflected off the cornea.
When his advisor at Michigan showed him a huge dataset collected from participants who were paid to watch television commercials, "ideas started popping up," he recalls. "The eye tracker measures attention, and due to its scarcity, attention has become more and more important to understand."
No one feels that need more urgently, however, than businesses that pay increasing amounts for television ads yet reach diminishing audiences.
Theories abound as to the most effective strategy for crafting a TV commercial—where and how often to place a brand in the ad frame. Some suggest using small, nonintrusive instances, while others recommend the hard-sell approach. There are also ideas about placing the brand early in the commercial, late (the so-called mystery approach), or early and late.
"The days when you could tell a consumer what to do are long gone," says Teixeira. "In the 1960s, the brand was onscreen all the time with a direct message: 'Drink a Coke,' for instance. Today, people are searching for valuable information that is relevant to them. They also want to be entertained, and the 'hard sell' that turns them off can be at the level of simply presenting the brand's logo for more than a few seconds."
Mystery ads try to entertain and draw people in by getting them to puzzle through what the commercial is selling, only revealing the answer at the end. But if a viewer watches the ad for 15 seconds before zapping it, the ad (and the sometimes millions of dollars spent to produce and place it) is completely ineffective.
"The dilemma is that our findings show that brand images cause people to zap," Teixeira says. "But they're a necessary evil; without the brand, viewers can't identify what is being sold. So how do you make an ad that includes the brand without causing a high level of zapping?"
In his research, Teixeira controlled for factors that could influence zapping, such as brand familiarity, the ad's visual complexity (too much or too little shows higher levels of zapping), and product category (perhaps you'd be more likely to watch a commercial for a "hedonic" product like chocolate than one selling a utilitarian item such as laundry detergent).
Teixeira isolated four variables that influence commercial avoidance: the presence of a brand in the frame, its size, how often it appears, and for how long.
Pulsing pauses zapping
Taken alone, brand presence automatically increases commercial avoidance. But Teixeira's research shows that changing the pattern of brand exposure (without decreasing the total amount of time the brand is shown) can lower zapping rates, and that a "pulsing" strategy in which the brand is inserted briefly and intermittently throughout the commercial is most effective, resulting in an average decreased zapping of 8 to 10 percent.
"If the brand is woven into the story line—not too overt, not too in-your-face—it's more likely that the consumer will get used to it and not have the urge to zap," Teixeira reports. "A conjectured byproduct of this research that may have a larger implication is in the delivery of any sort of undesirable news that an audience doesn't really want to hear. If you deliver it all at once, a person will tune out mentally. A better strategy would be to deliver the information bit by bit and intersperse it with other, more appealing items."
The eye-tracking technology also proves useful in showing the point at which viewers' attention begins to drift.
"We found that measuring the degree of dispersion in eye-movement patterns was directly related to distraction, and that the farther away a person's gaze is from that of other viewers, the more likely this person is to zap the commercial," Teixeira says. "We think that's something that can be used very efficiently—advertisers can see where they begin to lose people. Attention, as any other scarce resource, should be managed."
From a managers' perspective, altering the commercials to mimic a pulsing strategy is a virtually cost-free fix for a significant payoff, with zapping rates for some commercials reduced by as much as 25 percent in a lab experiment. If a company is paying to advertise on prime-time television during a show with an estimated 5 million viewers, 50 to 60 percent of those viewers (2.5 to 3 million people) can be lost with usual zapping rates; as many as 1 million could be recaptured by using the pulsing strategy.
"This isn't a constraint for the creative talents of advertisers," Teixeira says. "It's a generalized finding that companies across the board can benefit from this communication strategy of pulsing. And factors such as irritation and likability remain constant, so it doesn't solve one problem at the expense of another."
In a future paper, Teixeira will go deeper into evaluating the emotions of viewers.
"To make this analysis even more useful, we need to get a measure of the enjoyment or the feeling a person is experiencing while watching a commercial—and we want that to be as unobtrusive as the eye tracker, at the same level of microsecond measures." To do this, Teixeira uses an eye-tracking setup that also films a viewer's face: Software detects facial deviations and analyzes universal facial expressions for joy, disgust, sadness, anger, surprise, and fear. "This software enables us to understand the likelihood of each person feeling these six basic emotions so that we can understand on a moment-to-moment basis what a viewer is looking at as well as what they're feeling at any given moment," Teixeira says. "It's a way to measure attention and emotions as drivers of consumer behavior."
Will these findings influence how advertisers craft their ads in the future? Teixeira notes that there is already some evidence of pulsing in ads, as in the award-winning "The Happiness Factory" for Coca-Cola, and various automobile commercials that briefly show the brand logo of a car from various angles as it maneuvers a winding road.
"Sometimes advertisers say that they're creative and can't be constrained by science or academic research," he says. "But this isn't really a constraint, and our results show that it can help advertisers achieve the goal of getting their message out to more people."
In today's image-packed world, reaching a few more jaded eyeballs is no small thing.
About the author
Julia Hanna is associate editor of the HBS Alumni Bulletin.