New research from Wharton has uncovered some important differences in how shoppers use voice-assist devices such as Amazon’s Alexa and Google Home, and the findings could help retailers better target this growing segment of consumers.

In their paper titled “Consumer Preferences and Firm Technology Choice,” marketing professors Pinar Yildirim and John Zhang at Wharton, and Yi Liu at the University of Wisconsin-Madison compare consumers’ device preferences with their brand preferences, and they use that correlation to build a model of distribution and pricing strategy.

Nearly half of the U.S. population has a voice-assist device, although not everyone relies on it to shop. The professors said that’s expected to change as people get more comfortable with the technology and adoption increases, so firms need to be ready.

“We make a lot of decisions based on technology, and nowadays shopping is not an exception,” Zhang said. “We know that every consumer has a different brand loyalty. If you put a layer of technology on top of that, the question becomes how does the technology preference interact with the brand preference to play a role in our choices of products?”

To answer that question, the professors first divided the devices into two categories of function: those that offer shopping support by providing product descriptions and recommendations, and those that offer convenience by making ordering as easy as possible. A series of experiments yielded three key findings:

  • Consumers with stronger brand preferences who buy the same products over and over aren’t interested in devices that offer a lot of shopping support. They already know what they want, so they don’t care to learn about a product’s features or get curated recommendations.
  • Consumers with weaker brand preferences derive more use out of devices that offer shopping support to help them explore products.
  • Consumers want devices that give them ordering convenience, regardless of the strength of their brand preference.

“The findings are not that surprising if you break down the costs and benefits that these devices provide,” said Liu, who worked on the research while earning his doctorate at Wharton. “For convenience, the benefit is somewhat uniform for everybody. For information or assistance, who will this benefit the most? People who aren’t sure about what they want to buy.”

“We make a lot of decisions based on technology, and nowadays shopping is not an exception.”— John Zhang

Price Discrimination as a Function of Preference

Ordering convenience is prized among consumers, which leads the professors to a conclusion about firms that want to adopt tech-enabled devices: Make it easy to use.

However, with ease of use comes a loss of price discrimination. If ordering convenience is uniform across retailers and across devices, there isn’t much incentive for shoppers to choose one over the other. That’s where shopping support comes in.

“Once you adopt this technology, your ability to price discriminate will be different and the competitive space might also start to look a little different,” Yildirim said. “If the main benefit of the device is shopping assistance, that gives an opportunity for retailers to mitigate the competition by creating a difference in selling and the consumer’s willingness to pay.”

It’s well-documented that environment influences a shopper’s willingness to pay. For example, a customer who enjoys visiting a nicely appointed department store is willing to pay more for the same pair of jeans they could find for less at a discount retailer, just because of the experience. Zhang said the willingness-to-pay principle also matters when shopping is virtual, even though the parameters shift a bit.

“A very robust conclusion coming from our paper is that as a retailer, you always want to keep up with order convenience technology. If you don’t, it’s going to hurt you. And if you adopt decision-aide technology, you better make sure you are doing it better than anyone else,” he said. “With order convenience, you’re going to have a race in speed of adoption. With support, there’s a race in quality.”

Third-party technology providers are also part of the relationship between consumer and retailer, and they have to make decisions about how much they invest in support or convenience features. The professors said tech providers may prefer not to offer the highest ordering-convenience technology to all retail customers in order to spur competition for it and charge higher fees. In other words, the provider can maximize profits by holding back.

“Once you adopt this technology, your ability to price discriminate will be different….”— Pinar Yildirim

“The developer is in the position to induce competitive or monopolistic use of its technology. When [support functionality] is limited, the provider may stand to benefit from a monopolistic adoption of its technology, which it can do by under-providing its [ordering convenience functionality],” they wrote in the paper.

The Future of Consumer Technology

The professors said their research is the first of its kind to examine consumer preferences for technology-enabled devices based on brand loyalty, and build a model for managers to price and sell across these devices and traditional channels. They expect more research will follow because technology is evolving rapidly.

Yildirim joked that most of her students are now using Alexa to get the weather forecast and turn off the lights, but who knows what those devices will be able to do in a few more years. Rapid advancements in generative AI make the research even more compelling.

“There is still fairly little understanding of how consumers are interacting with AI and how that shapes their decision-making, and in turn how the software learns from the response,” she said.

Liu and Zhang agreed, saying the interplay of consumer psychology and marketing offers endless research possibilities. They pointed to the four components they studied in their research: consumer brand preference; consumer technology preference; firms’ adoption and use of technology; and consumer choice.

“I think this paper is a modest beginning of looking into the complex interactions of those elements,” Zhang said. “Going into the future, any one of those elements could change. If we follow the path of this particular research, we will have a long road ahead for sure!”