Auto insurance companies are some of the largest advertisers, with top brands spending upward of $1.5 billion per year, even though they’re household names. When Geico runs another spot featuring its ubiquitous gecko mascot, or Progressive cranks out another Flo ad, are they wasting their money?
“That’s where the puzzle starts,” says Navdeep Sahni, a professor of marketing at Stanford Graduate School of Business. After a century of research on advertising, scholars still don’t have an empirically solid grasp of exactly how or why it works. One leading theory suggests that advertisements provide consumers with new information. If that’s the case, why do the most prominent brands need to advertise? Another influential theory suggests that advertisements create associations in consumers’ minds, hitching a particular brand to a category.
“But if you push that thinking further, it doesn’t actually tell you what to expect,” Sahni says. “If you’re primed to think about Geico, that might also make you think about Progressive because the two things are associated. Or it might make you think about Progressive less. The theory isn’t diagnostic or precise in what it says.”
Sahni and doctoral candidate Yifan Yang go a long way toward resolving these issues in a new working paper that shows the dramatic benefit of auto insurance advertising. They found that ads not only boost visits to an advertiser’s website but also interfere with consumers’ recall of alternative brands, dislodging the competition from their minds. The findings help explain why spending millions on repetitive campaigns, even by well-known brands, is essential to remaining in people’s memories.
The Persistence of Memory
The research centers on a large-scale randomized controlled trial, in itself a valuable contribution to the field. “Ad measurement has been a struggle for a very long time,” Sahni says. Most investigations have relied on either lab experiments and surveys, which may not accurately mirror the real world, or comparisons between markets, an approach that may be confounded by looking at ads that target different groups.
Sahni and Yang’s experiment ran on a single day on one of the world’s top five most visited websites. Selecting more than 325,000 visitors who had installed a related browser plug-in, they showed 75% of them a banner ad for an auto insurance company. The other 25% saw a public service announcement unrelated to insurance.
Counterintuitively, the competition’s previous marketing efforts actually boosted the power of the advertisement on the day it was displayed.
The researchers observed a surge in visits to the insurer’s website on the day of the experiment: The likelihood that individuals visited the site increased from 0.2% to 0.8%, a 300% increase. The researchers continued to track users’ online behavior for three more months, and although the effects of the advertisement decreased with time, the total effect was equal in magnitude to the effect on the day the ad ran.
“This very strong effect dissipates in a matter of days,” Sahni says. “However, if we continue to look into the future, then the effect comes back, and the total cumulative future effect of advertising is the same as what we see on the day of the advertisement.”
Within their sample, the researchers were also able to see what effect advertising had on the insurer’s competitors. This, Sahni notes, “really helps us understand what’s going on under the hood, what the mechanism is that’s driving these effects.”
“A Combative Environment”
The researchers also looked at the effect of competitors’ ads on consumer behavior. Counterintuitively, in markets where competitors advertised on television in the weeks leading up to the experiment, the effect of the website banner ad was more pronounced: The competition’s previous marketing efforts actually boosted the power of the advertisement on the day it was displayed.
This occurs because the advertiser being studied and its competitors are locked in a running battle for consumers’ attention. Ads from the competition pushed the advertiser from consumers’ minds. In turn, by running a banner ad, the advertiser dislodged the competitor and regained a prominent association with auto insurance. (If the competitor hadn’t previously advertised, then the advertiser may have already been top-of-mind.) “This creates an incentive to advertise more, which is one of the key implications of our work,” Sahni says. “This is a combative environment.”
With these effects quantified — the strength of advertisements when they’re observed, their persistence in consumer memory, how they interact with competitive advertising — the researchers developed a predictive model that they tested on novel data. Their new model outperformed the standard models used widely in the ad industry.
It’s not clear if Sahni and Yang’s results might hold outside the context of auto insurance, but Sahni says this work is a step toward understanding advertising more broadly. “The objective of this paper was to lay the groundwork for developing these more detailed models,” he says. “Our paper stands out in predicting effects, and it gives us a path for thinking in a more structured and rigorous manner about consumer perceptions, about what is top-of-mind when they’re making decisions.”
