Women’s clothing retailer Bebe is rewiring the way it turns data into personalized offers for young, fashion-forward females on their devices of choice.
Founded in 1998 and now operating more than 200 retail stores with 100 international franchisees, the company hired a chief digital officer, Erik Lautier, in January 2014 to improve conversions on Bebe’s ecommerce site.
Initially, Lautier tackled the operational basics – things like blocking and tackling, A/B testing and developing on-site product recommendations – and it selected tag-management and ecommerce personalization platform Qubit last June to lay the groundwork for more personalization across its desktop site and mobile.
“Increasingly, we’ve started to tie the site experience to offline using Qubit to determine that when somebody clicks on a marketing banner and lands on the site, did we show them a product that is relevant to the first few actions they took?” Lautier said.
But Bebe is also leveraging Qubit to help tailor promotions for users who hit their site and who are, for instance, within a 50-mile radius of a new Miami retail location by geotargeting nearby users.
It’s using Qubit’s new platform called Visitor Cloud, a database built by a handful of Google alums who tout its ability to process 60 billion-100 billion marketing events per month at breakneck speed because it’s a single platform built from the ground up, not a series of “bolted-on solutions,” as Graham Cooke, Qubit’s CEO, described.
Visitor Cloud is generally integrated with users’ own ecommerce, business intelligence or email or CRM systems, or it can be paired with personalization engine Qubit Deliver for campaign execution.
By using Qubit’s Visitor Cloud, a key part of its marketing mix as well as a baseline for its data infrastructure and tags, Bebe improved desktop conversions by 18% year over year and mobile conversions by 93%.
“We’re no longer A/B testing something for the site – we’re A/B testing things down to the person and their preferences and we want to give the individual items they’d like to see, rather than what the aggregate wants to see,” Lautier said.
One example was Bebe’s investment in video. Pre-Qubit, the brand deployed 8 megabytes of lifestyle videos on the home page at its creative agency’s suggestion, anticipating it would inspire customers to buy.
But once Bebe ran tests, it discovered the lifestyle videos were not contributing to overall conversions, and it instead slotted in product recommendations tailored to the individual shopper in the place of sight, sound and motion space hogs.
“We are also increasingly turning our attention to mobile in terms of conversion efforts and our overall hope is Qubit helps us personalize in a much better way cross-channel,” Lautier said. Bebe crossed the 50% mark in mobile traffic this last holiday season. “In some ways, it introduces new challenges because an increasing proportion of traffic is now coming through a channel that converts worse than desktop.”
Although Bebe had a noticeable uptick in mobile conversions this year compared to last, like many retailers, it had to avoid repackaging desktop functionalities for mobile.
“There is still a little bit of a hurdle with getting people to transact on mobile,” Lautier added. “They’re fine reading the email there, clicking through to browse maybe and then they still typically want to transact on the desktop and buy.”
Thus, it’s resorted to mobile-specific promotions for a certain percentage off when a user spends $250 on mobile.
“We want to encourage more people through that hurdle of converting because our mobile site is actually quite good and performs very well for us when it used to be a bit of a disaster for us,” he said.
Although Qubit has helped Bebe create that coveted single view of the customer, Lautier said the next step is attaining more granularity in mapping the customer journey.
If Bebe buys a nonbranded keyword like “party dresses,” and a user clicks through on that search term vs. display or email, “we’re not able to say yet that SEM drove that ultimate conversion,” he said. “The data is there, and we recognize the importance, but we don’t have the tools or the personnel yet to really tie it together.”
