Sales Can Stagnate if Online Retailer Promotions Are Limited to Previous Purchase Patterns

By Jeanne Leong

To induce shoppers to purchase more items, online retailers often offer customized recommendations to customers based on prior purchases, but a Rutgers University‒Camden researcher says that approach could do more harm to the company than helping to increase overall sales.

A study led by Nathan Fong that appears in the Journal of Marketing Research shows that customers who receive these highly personalized, targeted promotions are more likely to purchase from those same categories, and less likely to look at other types of products or buy other items.

Fong, a Rutgers‒Camden assistant professor of marketing, and his co-authors conducted field experiments by having a popular e-book mobile company based in Asia send book promotions to nearly 40,000 customers through the company app’s push notifications.

As one of the top five e-book platforms in Asia, the company serves a large user base with 130 million visitors to its app every month, where customers can access genres such as biographies, fantasy, nonfiction, romance, and science fiction.

Nathan Fong, a Rutgers‒Camden assistant professor of marketing

Nathan Fong, a Rutgers‒Camden assistant professor of marketing

In the highly targeted promotions, the company sent customers offers for books of the same genre that they previously purchased, and offered three books of different genres. The customers could download and read several chapters for free.

The researchers found that consumers who received a targeted promotion in a genre of books that they purchased earlier chose the targeted promotion item 2.5 times more than the untargeted item.

While the rate of customers’ direct response to the same-genre promotion is higher than to the different-genre promotion, Fong says that firms need to be aware of the opportunity costs of running such promotions.

The researchers say that for every 100 customers who received offers selected to maximize response to the promotional offers, the company is losing out on 65 downloads when they only focus on maximizing downloads of the same-genre books they are promoting to customers.

“You can’t rely on a single key performance indicator,” says Fong. “You’ve got to look at a variety of metrics to make sure that when you are optimizing on this one metric, that it’s not actually hurting you elsewhere.”

Fong says managers do not always consider what a customer would do if the retailer suggested different items, or had not sent any promotions.

“Maybe they would have gotten into historical fiction, as well,” says Fong. “Or maybe young adult fiction, or something else.”

The researchers propose that firms that use highly targeted promotions should pay close attention to customer search patterns and the diversity of sales. “They play the odds from the information they have,” explains Fong. “Often, when you don’t have a lot to go on, the way to maximize response is to promote more of the same.”

Trying to sell customers similar products to those they’ve already purchased can be successful, Fong says, when customers want to buy the same types of items, such as phones and other electronics, which some customers want to upgrade every year or two.

The researchers say broader search patterns can generate cross-selling opportunities and diversity of sales, and help the company retain customers.

Fong says companies such as Netflix take a more nuanced approach to targeted marketing by promoting items to customers while also learning more about their preferences. If a customer watches certain genres or sub-genres, Netflix offers programs in that same category, and at the same time, they select random programs for a customer to consider.

“They’ll try to mix it up because they might find that you like this other type of programming as well, and they can get you interested in watching a whole other category, which gets you using the service more,” says the Rutgers‒Camden researcher.

Fong’s co-authors of “Targeted Promotions on an E-Book Platform: Crowding Out, Heterogeneity, and Opportunity Costs” are Xueming Luo of Temple University, Xiaoyi Wang of Zhejiang University’s Yuquan Campus, and Yuchi Zhang of Santa Clara University.

Comments are closed.