How fashion and beauty brands are using AI to drive mobile purchases

It’s what happened at DvF.com. “Mobile is at least 50 percent of our traffic, but not our revenue. To be honest, we all kind of gave up on fixing that,” said Felipe Araujo, the head of e-commerce at Diane von Furstenberg.

That was until a major rebrand last year, following the departure of Diane von Furstenberg as the creative director of her eponymous brand, when the DvF team overhauled its e-commerce site. With a new logo and designer, Jonathan Saunders, the brand needed a smarter online shopping experience to bolster the reboot. With a new site, Araujo prioritized improving product discovery and navigation on mobile, in order to turn that traffic into sales.

Working with personalization platform Qubit, DvF set up an algorithm-based discovery feed on its mobile site that’s curated using artificial intelligence and tailored to each user. Through data culled by the brand and Qubit, the feed reflects an assortment that changes, depending on how a visitor has shopped in the past.

The goal is to use AI to build a personalized product feed that will make the discovery process on the mobile web easier. Today, Fashion brands want to not just be mobile-first, but mobile-only. But there are several barriers mentioned by retailers when explaining the stalled conversion rates on mobile: finicky checkout fields, small screens, the inability to get customers to download and use a mobile app, and the friction involved with driving sales from social media, among them.

“This solves a few mobile pain points: You’re creating an app-like experience in real time, while making a custom journey for each user. Thats where the opportunity is,” said Bud Goswami, lead data scientist at Qubit. “When you’ve paid to have all that traffic on the site, you need to give people a reason to stick around longer.”

For instance, on the DvF feed, complementary shoes, accessories and other items will be shown to match a previously purchased pair of pants, while everything in the feed will reflect sizes and styles similar to the customer’s preferences, based off of machine learning techniques. If a customer hasn’t visited or purchased in the past, the feed will show top-trending items and items near selling out, along with the exact count of how often an item has been clicked on. The feed will also change depending on factors like geographic location and where the customer clicked through to the site from, like an email newsletter or Instagram.

“I was very hesitant to show the highest ranking products on the site because, in the middle of a rebrand, we wanted to be the authority, talking to our customer and showing her what to look at,” said Araujo. “But we needed a discovery element that would help them through the journey. We have eight product categories on the site and 300 new products a season. It’s hard to go through all of that on mobile.”

Instagram, in particular, has proven to be a black box when it comes to driving direct sales through brand posts. Even though the platform added shopping tags last year, sending users to brand sites directly from posts for the first time, the results were disappointing, said Araujo. On an Instagram post that got 4,000 likes, for instance, about 100 people actually clicked through to see product details. Only about 10 people then clicked through to the actual site.

“We’re attempting to bring that discovery phase onto the site instead, so it feels very in line with what you’re seeing on social apps,” said Araujo. “That transition between discovering and shopping needs to be a lot easier.”

ColourPop, a cosmetics brand with a cult following of 5.1 million on Instagram, is also using the AI discovery feed to bridge the gap between its social media fans and mobile checkout. Rather than just recreating a shoppable Instagram feed on its site, like brands including Nike and third-party platforms like Curalate have done, the brand wants to promote purchases by putting the right products in front of people through an algorithm-based feed.

A desktop version of the discovery page

“The biggest thing about mobile shopping right now is if you don’t know what you’re looking for, discovering that is really difficult,” said Nate Dierks, ColourPop’s director of technology. “This has given me a lot of thought about the purpose of the mobile website: It needs to be just a stripped-down discovery page that shows you what other people are looking at, what’s popular, what’s about to sell out, and what you should know about this brand.”

According to ColourPop, the brand has gotten a 31 percent open rate on the discover feed icon since rolling it out in beta, and a 4 percent uplift in revenue per visitor.

Dierks said brands have to come to terms with the fact that product discovery on mobile is happening “outside of the brand’s four walls,” and make proper adjustments. While a mobile app was once thought of as the way for brands to carve out undivided attention on mobile, building a use case for a specialized app has proven difficult.

“Brands won’t be building out new apps anytime soon. We’re not going to do it,” said Dierks. “But if you can experience a brand through other apps like Instagram or YouTube, then go and shop a feed that knows not only if you came from an app but what people like you are shopping, then people will shop.”