The apparel retailer boosted sales in 2018 as a result of its redesign efforts, which include a trending feed and recently viewed nudge when a shopper returns to its website.

Women’s apparel retailer Diane von Furstenberg, known for its wrap-style dresses, launched a new e-commerce website about a year and a half ago after it struggled with a site that was too generic and didn’t work well on mobile.

The retailer redesigned its e-commerce site “with the right team focused on personalization,” Felipe Araujo, senior director of e-commerce at Diane von Furstenberg, told attendees during the session “Personalization: The Site Changes That Drive Meaningful Sales” at the 2018 Internet Retailer Conference & Exhibition in Chicago.

Diane von Furstenberg, No. 836 in the Internet Retailer 2018 Top 1000, has boosted its sales so far in 2018 compared with 2017, Araujo said, crediting much of that revenue increase to its redesign efforts.

The apparel retailer first removed some of the excess technology that it felt was bogging down the website—such as reviews and content titles on product listing pages—and reduced the number of product filtering options. It also added a number of personalization elements with the help of personalization platform provider Qubit and its Aura program.


When a shopper visits the homepage, she sees a feed of items that are “trending now,” which is a lot like product recommendations, and shows consumers items that have been recently purchased by other customers and other popular items on the site. The trending feed is automated for most visitors, including past customers, new visitors and returning non-purchasers, who are shoppers who have browsed items, but did not purchase anything.

The personalization elements also adapt to what the shopper is looking at when on mobile. If someone clicks on items in the sales category and then clicks on the wrap dresses category, the machine-learning algorithm implemented in the website will re-sort the category navigation, changing it to show wrap dresses and sales items since that is what the shopper is most interested in viewing, Araujo said.

Diane von Furstenberg saw a four-times increase in conversion rate for people who interacted with Qubit’s Aura versus regular mobile users.

In addition, if a shopper returns to after browsing a few items on her last visit, she will see a pop-up message, or a nudge as the retailer calls it, that shows her items she recently viewed, as well as the trending feed and an e-mail sign-up to become a “DVF insider.” The recently viewed nudge accounted for $125,000 of its revenue, Araujo said. “We’re trying to jumpstart the experience, to get you right back where you started searching,” he said.


The retailer also found engagement and conversion on its mobile site to be an ongoing issue before it redesigned its website.

A Qubit study of 2 billion interactions and 120 million purchases—although the company did not specify the time frame of this study—showed that 28% fewer shoppers add products to their bag on mobile and 44% fewer shoppers convert on mobile compared with desktop shopping. Diane von Furstenberg cited reasons for its shoppers not buying more on mobile, such as the desktop site was faster to browse, easier to find items on and easier to make a payment through, Araujo said.

But since implementing changes and making its website more mobile friendly, Diane von Furstenberg’s mobile conversion rate increased 400%, Araujo said.

The initial iteration for the website redesign process took about three weeks to ensure the personalization elements worked, he added. But now that the retailer is one-and-a-half years into making these changes, “it has unlocked infinite possibilities,” Araujo said.


“We will continue to iterate,” he went on. “You don’t setup an experience that’s working great and move on. You come back and start to optimize it because that’s when it becomes really fun.”

In an ongoing effort to understand the most vital business strategies of merchants that sell online, Internet Retailer is taking a deep dive on conversion rate. Help us improve our estimates and category benchmarks by answering the two questions below.