If the passenger sitting next to you on a flight buys a snack or a movie, does that make it more likely that you, too, will buy something?
Yes, says Stanford Graduate School of Business professor Pedro M. Gardete. He examined what’s known as social effect—when our behavior is influenced by the behavior of others—by studying one market: items for sale on an airplane during a flight.
Our friends and peers, Gardete says, have plenty of sway over what we buy and when we buy it, but there is still much that marketers and advertisers don’t know about the magnitude of that influence. His new research found that when we see someone near us make a purchase, we’re 30% more likely to buy something ourselves, and the relevance of that goes far beyond the confines of a plane’s cabin, to marketing strategies in a wide array of industries.
Gardete chose the in-flight setting because it is a confined environment, with a variety of items for sale and an abundance of recorded transaction information. In-flight sales fall under the “ancillary revenue” category for airlines, which have seen major growth worldwide. From 1995 to 2013, ancillary revenue—sources such as food, beverages, and in-flight entertainment—grew about 26.5% per year for US airlines, more than five times the airlines’ total revenue growth over the same period. Gardete examined the purchase data of 2,000 flights from a major US airline during January and February 2012. That sample included about 257,000 people who together performed 65,525 purchases, an average of 33 per flight. Because the transactions were made with via credit card, Gardete had precise information about the transactions, including the buyers’ flight numbers, seat numbers, what they bought, and what time they bought it.
In order to measure the social effects of purchases by one passenger on other passengers, Gardete set up a quasi-experiment. He looked at consumers in pairs, with each person sitting in the same kind of seat—aisle, window, or middle—and one right in front of the other. He assumed paired passengers were fairly similar because they were seated near each other, in the same type of seat and flying economy on the same flight. One passenger in the pair was in the treatment group, meaning the person observed a purchase by the passenger seated next to him or her. The other person in the pair, the control, was seated in the same kind of seat one row ahead but did not observe a purchase.
Gardete found that passengers were a lot more apt to buy something if someone next to them made a purchase first; in fact, they were 30% more likely to buy after exposure to a purchase than they were before exposure. That did not hold true if they saw a purchase by someone sitting behind them or diagonally in front of them, says Gardete. “It’s only the person next to them that affected them,” he says.
If two strangers could have this much influence over each other, Gardete wondered what would happen with two friends. Those friends, he reasoned, likely share the same tastes because of homophily, which is the tendency people have of associating and bonding with those like them. To test that, he analyzed the purchasing behavior of people traveling together under the same reservation number, assuming they knew each other. He found the likelihood that a passenger will buy doubles if the person who is next to them making a purchase is someone they know. “This could be because they share the same tastes or because they are influencing each other,” Gardete says.
These findings show not only that people pay attention to what their friends buy, but also, and perhaps just as important, that someone who has already bought a particular item—before seeing anyone else buy it—is even more sensitive to social influence than someone who hasn’t yet made a purchase. “Even though they’ve already bought something to eat or watched the movie, they remain extremely open to social influence, more so than someone who has bought only from social influence in the first place,” Gardete says. “When people buy a product, they aren’t just signaling they like the product or the experience, but they are telling you they are also more sensitive to promotions and very sensitive to social influence. That’s something that hadn’t been found before.”
Gardete’s findings have important implications for marketers, who are looking for information that tells them how consumers are influenced by their friends and the particular products and services their friends promote. It builds on a shift already taking place in digital advertising toward a reliance on social graphs, where businesses use a consumer’s social network connections to market to them. “This is the bread and butter of marketing companies right now, as they try to find the best way to approach consumers,” Gardete says.
And the best way may be through a consumer’s peer group. If a company sells bicycles, for instance, and you’ve bought an expensive bike from them, they know you’re interested in biking and it’s likely that some of your friends are too. That’s where the effectiveness of social targeting comes in, says Gardete. As his in-flight research shows, our peers can have a significant influence over whether and what we buy.
The next step, says Gardete, is to look at ways companies can take advantage of social effects. “We know now people are prone to social influence. I’ve documented the features that are necessary for value to exist for companies, but how do they take advantage of that?” he asks. Airlines, for instance, might want to reward its passengers who have made in-flight purchases by sending them vouchers before their next flight or even right after they make a purchase, in real time, says Gardete. “Implementation is a whole side of this that we don’t know much about yet. But it will be an interesting problem to tackle.”