Farfetch identified new e-commerce site visitors who were highly engaged with its site and displayed to them a page overlay with information about its shopping, shipping and returns policies. The retailer also ran campaigns for return shoppers who hadn’t purchased before and for all new visitors based on analytics from vendor Qubit.
United Kingdom-based online fashion retailer Farfetch sells high-end clothes for men and women from designer boutiques in 24 countries, with single items priced from the low hundreds of dollars to $3,000 or more. Customers arrive on the site from more than 170 countries, and convincing them to make such expensive purchases on the web requires building trust in the retailer’s shopping, shipping and returns policies, says Kelly Kowal, digital marketing director.
With help from behavioral analytics company Qubit, the retailer found a way to do it. Farfetch ran A/B tests based on customer segments that Qubit identified by measuring site visitors’ behaviors. In the first, it displayed a screen overlay with information about its shopping and shipping policies to customers who were new, but also highly engaged with the site—they’d viewed five to 10 pages in one visit. Those shoppers’ conversion rates increased by 17%, Kowal says.
“I didn’t think it would have as high an uplift as it did,” she says. “It sounds so simple—just showing people more info about your policies.”
In a similar test, displaying information about the retailer’s policies to all new web site visitors, conversion rates went up by 7%, Kowal says. That test displayed the information in a different type of pop-up overlay known as a lightbox, which doesn’t cover the whole screen. In addition to explaining the retailer’s global shipping and returns policies, the overlays in both tests also used a customer’s location to tell her whether the prices she’s viewing include country-specific taxes or duties, Kowal says.
Farfetch is now testing a way to help customers who seem hesitant or confused, which Qubit identifies by measuring how long they hover over a product listing. After a certain amount of time spent hovering, the retailer displays a message to those shoppers suggesting that they “shop our edits,” or try browsing through one of its men’s or women’s collections.
“We have a lot of products—it can be overwhelming if you’re new,” Kowal says. “So for new users who we can tell are engaged but not really finding what they’re looking for, we serve them an overlay asking if they need inspiration, which then takes them to a part of the site that’s a bit more curated.”
So far, adding these behavior-triggered messages has increased traffic to the edits pages by 14.3 times, which is in turn boosting conversions, she says, without giving the exact rate, as the test is still going. In general, though, she says customers who visit the shopping edits on the site convert at a higher rate than those who do not.
Qubit provides a web-based dashboard for retailers to measure customers’ site behavior and segment them into groups for A/B testing. However, Farfetch mainly relies on its weekly calls and communications with its Qubit account managers to get the custom reporting it needs and to discuss strategies for running new tests on the web site, Kowal says. “A lot of the analysis we ask for is bespoke and they’re very good at delivering exactly what we ask for,” she says, adding that the web dashboard does provide a tool that her team can use to query on its own when it chooses to. “They’ll even give us the raw data if we want to do the analysis ourselves or cross-check it.”
Farfetch pays a flat monthly fee to use Qubit’s web site analytics and A/B testing technologies, including all support, Kowal says, declining to disclose the exact price. However, with the amount of testing the retailer runs—it launches a new test every few weeks—and the speed with which it reaches conclusions about whether site changes are effective or not, she says the service is “extraordinarily cost-effective.”
Farfetch.com is No. 188 in the 2013 Europe 500 Guide.