Unnati Narang has a longstanding interest in retail. In India, where she grew up, she co-founded a startup that sold musical instruments online and offline. Her interest in retail and digital technologies  eventually led her to a PhD program in the US and a job as an assistant professor at Gies College of Business at the University of Illinois Urbana-Champaign.

While she was working on her doctoral dissertation at Texas A&M, Narang began studying issues relevant to omnichannel marketing, empirically analyzing data for a large US-based retailer of consumer electronics and wireless services. The chain is primarily brick-and-mortar — it has 4,000 stores — but it also has an ecommerce site and a mobile app. Online sales comprise about 10 percent of the chain’s overall revenue, though that number has been growing over the past five years. Narang was especially curious about how the mobile app affected sales across the board.

In the first part of her dissertation, Narang established that a retailer’s app leads shoppers to buy more through the retailer’s website and stores as well. She showed in a 2019 paper that the app assists shopping in other channels. 

“Apps make it easier for shoppers to find more information about products and to locate stores near them,” Narang said. “There is an advantage to the app that makes shopping in stores much more convenient.”

However, she also noticed that even though app users bought more merchandise than other customers, they also returned more unwanted items, which made her even more curious about how the app affected their shopping habits.

Impact of App Failures on Multi-Channel Shopping Behavior

While examining the retailer’s data, Narang noticed that there were periods when the app failed due to server errors in the back end. She realized this would be a perfect opportunity to observe what customers did when they were suddenly deprived of this one channel for shopping. Would they switch to another channel, like the website, or go to the store in person? And how this would affect overall revenue? 

Narang and her research partners, Venkatesh Shankar of Southern Methodist University and Sridhar Narayanan of Stanford, studied one particular outage on April 11, 2018, when the app became unavailable for two hours, between 12 and 2 pm CT.

On the day of the outage, 136,689 users tried to access the app, and nearly half — 70,568 — failed. The researchers compared those who experienced the failure with those who did not over a two week period before and after the failure, and found a 7% decrease in purchases. In the short term, this cost the company $1.08 million in revenue and, in the long term, an additional potential $1.89 million due to lost customers, i.e., shoppers who permanently reduce their spending with the retailer over the long term.  The retailer also saw a decrease in discounted purchases over the two-week period because customers were unable to access special deals that were offered through the app.

The most surprising finding, though, was that the app outage didn’t affect online shopping. This is due to "channel switching," Narang says. When the app became unavailable, customers would simply switch to the website, especially if they had used the website before. But what the researchers didn’t expect was that the app outage would lead to a decline in in-store shopping.

“The results we found supported the idea that the app serves as a way to search,” Narang said. “And when that search breaks down, shoppers don’t really want to go through the extra effort to go to the store and find the product that they want. So the pathway from app to store gets shut down because this ability to search for information is no longer there.”

The paper, Cross-Channel Effects of Failure in a Retailer’s Mobile App, was recently published in The Journal of Marketing Research. Narang has presented the findings at various conferences to a favorable response.

“I think what people find really cool is the ability to observe what we observe,” she said. “Because if you look at a lot of the traditional marketing work on service failure or customer reactions to failures, it is very much based on a survey or qualitative data. We were one of the first researchers who had this ability to observe across the board what customers were doing and how they were impacted by app failure. Ours was the first study that could quantify how much an app failure would affect actual purchases and spending in the future, in what channels, and why.”

This will be important in the future, she believes, as shopping, even at brick-and-mortar stores, becomes more app-focused. Retailers use these apps to collect more information about their customers’ shopping habits – not just what they buy, but where and when they buy, and other places they spend their time and money. This, of course, also raises more questions about privacy and government regulations.

“We’ve been noticing a lot of the tech products are not as robust as we would like,” Narang said. And retailers should be paying close attention.