Like many new retailers, online lingerie seller Adore Me initially devoted a significant amount of resources to customer acquisition. The strategy worked: within six years of its founding, the members-only site, which offers new collections of merchandise every month, had grown to $100 million in annual revenue.
However, “at some point we realized that our revenue had shifted from new customers to our existing customers,” said Josselin Petit-Hoang, CRM Marketing Manager at Adore Me. But maximizing engagement and sales from those customers was going to be a challenge, because “our communications with these existing customers were very basic,” said Petit-Hoang in an interview with Retail TouchPoints. “We were only using email, and we were essentially sending the same message to all of them. But not all customers are equal: Some are very engaged; they buy a new set of lingerie every month. Others come to the site twice a year, buying one set for the spring and another for the fall.”
Adore Me decided to optimize communications by carefully segmenting its customer base with the help of the Optimove Customer Marketing Cloud solution. Initially, the retailer’s marketing team used Optimove to send daily emails to five distinct customer personas. With the help of the solution’s predictive customer modeling tools, Adore Me then identified additional personas that would respond to highly targeted campaigns.
Advertisement
In less than one year, Adore Me was communicating with 66 unique customer personas on a monthly basis, using email, mobile push notifications, in-app messages, Facebook Custom Audiences and the Google Display Network. The results have been dramatic:
• 15% increase in monthly revenue;
• 22% increase in average order amount;
• 2.3X growth in the number of active customers; and
• 85% of customer campaigns are fully automated.
Using Multiple Metrics To Segment Customers
Adore Me uses a wide range of metrics to define customer personas. “The Optimove software gave us the opportunity to segment based on attributes such as who customers are and where they live, but also web site behavior, order history and which device they are using to connect with us, for example a desktop or an iOS,” said Petit-Hoang.
Armed with this information, Adore Me has learned to adapt the content and delivery method of customer communications as well as the frequency. “The customers that are very engaged, the lingerie lovers, want to see and buy a lot of new styles,” said Petit-Hoang. “We’ll send them updates on what’s new and what’s just been released. Then there are people that buy only rarely. For those customers we might send them discounts because we want to try to re-engage them.”
Adore Me uses the Optimove solution to continuously test the effectiveness of customer communications. “Optimove allows users to treat each marketing campaign as an experiment,” said Pini Yakuel, CEO of Optimove in an interview with Retail TouchPoints. “When Josselin [Petit-Hoang] is running a campaign, he can measure the exact incremental dollar uplift on a weekly and monthly basis. That’s often difficult in marketing, because often different campaigns are stepping on each other’s toes. We want these to be statistically significant and ensure that every experiment is ‘clean’ in terms of data and results.”
The Importance Of A/B Testing
Because Adore Me has established a large number of customer microsegments, A/B testing helps the retailer yield more significant results. “When people think about an A/B test, it’s usually about comparing the color of a button or whether one product picture works better than another,” said Petit-Hoang. “But if I want to target our Active VIPs, I might send one group a message with a discount, via email and Facebook. Another group might get a push notification, with a free shipping offer. That’s two streams of variance that we can analyze.”
Adore Me also is employing strategies that maximize the benefits of these sophisticated A/B tests. “We’re not simply going to declare one method the ‘winner,’ because the ‘loser’ still appeals to some customer groups that prefer that method of communication and that offer,” said Petit-Hoang. “With the predictive modeling providing an infrastructure of microsegments, even the ‘losing’ action will serve best for a particular microsegment.
“Having accurate financial uplift results from every campaign makes it easy for us to know what’s working and what isn’t,” Petit-Hoang added. “This helps us optimize our efforts an on ongoing basis.”