In the past two years web sales for Cosabella.com have grown to account for 15% of the high-end lingerie retailer’s sales, up from 8%, says Courtney Connell, marketing director for the retailer and wholesaler.

The rest of the brand’s sales are mostly conducted via wholesale distributors, including at luxury department stores like Neiman Marcus, No. 36 in the Internet Retailer 2016 Top 500 Guide, and Nordstrom Inc. (No. 18).

Last year, the 34-year-old family owned merchant, which first began selling direct online in 1996, cut ties with the digital agency that was helping it with email marketing. It then began exploring using a vendor to help it better personalize email messages and boost online revenue from its list of about 50,000 email subscribers, Connell says. “We were collecting a lot of shopper data, but we didn’t have the capacity to act on it,” she says.

Cosabella in October signed a contract with Emarsys to help it customize the emails it sends to its subscribers three days a week (Mondays, Wednesdays and Fridays). After implementing Emarsys the following month, email revenue increased 60% year over year and email open rates increased about 4%. What’s more, Cosabella’s popular 12 Days of Cosabella holiday email and marketing campaign that contains a different offer or product for 12 days in December, generated 40% to 60% more sales in 2016 compared with a year earlier, Connell estimates. And for the first time, Cosabella did not offer discounts in that annual holiday email promotion, instead focusing on increasing editorial content, personalized suggestions and user-generated content in email messages to draw shoppers to the site and pique their interest.

“The new system allows us to leverage data and lot of great merchandising capabilities,” Connell says. “We can take consumer data we have and feed it to Emarsys and make it actionable.”

For example, Cosabella can target shoppers based on what messages they responded to in the past or items they previously purchased. If it notices a shopper is typically a low spender or has visited the site from emails, but has not yet bought, Cosabella might send her an email promoting a “happy hour” sale where she can nab 20% off from 6 p.m. until midnight, Connell says.

Creating emails through Emarsys also is much easier and faster than Cosabella’s old system. Before it took an employee skilled in HTML or graphics up to three hours to create an email. Now just about any employee can create an email in less than an hour because the template-based Emarsys system is easier to work with, Connell says. Emails today also contain much more content and the placement of the content changes based on what Emarsys and Cosabella know about the shopper.

For example, a message might contain three sections: new colors, new collections and older items that are on sale. Cosabella might heavily promote new colors and collections for bigger spenders and eliminate or minimize the discounted goods portion of the email. Cosabella also can easily change the call to action button depending on who is receiving the email.

Cosabella is adding much more content to its emails using Emarsys’ recommendations. For example, an email might contain blog content, user-generated content from Instagram, and new collections and new colors, whereas before the work required to create such an email typically limited message content to a single theme.

“We’ve been able to focus more on lifestyle,” Connell says. Case in point: the aforementioned 12 Days of Cosabella holiday campaign used to focus on a specific product, highlighting perhaps an underwire bra one day and a loungewear set another. This past year, Cosabella focused on themes such as “sexy,” “trendy” or “flirty,” and it experienced an uptick in sales.

Emarsys takes historical shopper data and then uses machine learning to identify shopper patterns and put shoppers into categories such as best shoppers, worst shoppers and other buckets based on what they buy. The vendor also identifies what it calls “hot defectors” or shoppers who open email and visit the site but have not bought recently and shoppers who are nearing inactive buying status, meaning they are approaching 62 days since they have last bought.

Using this data, Emarsys suggests the best message or offer for Cosabella to send, based on what is predicted to generate the most revenue. For example, rather than a retailer blasting out a 20% discount to its entire email subscriber base, Emarsys can suggest some shoppers get $5 off instead and others simply get a message highlighting new bralettes without a discount.

While the Cosabella team still takes these suggestions from Emarsys and decides which ones to act on, Emarsys is working with 10 retailers that give the decision-making authority to Emarsys about what an email or marketing campaign should contain for each shopper and the vendor then automatically creates the email. That service, which Emarsys calls AIM (Artificial Intelligence Marketing) helped weight-loss supplement retailer Evolution Slimming generate a 28.6% increase in revenue in a test when compared with a control group that received a default offer. Emarsys launched AIM in late 2016 shortly after it raised $22.3 million, bringing the company’s total funding to $55.3 million.

Cosabella says it plans implement AIM soon, along with other AI-based marketing services it uses. This includes user testing software from San Francisco-based Sentient Technologies that uses artificial intelligence to enable companies to test nearly an unlimited number elements across multiple pages at once to determine the arrangement that works best. It also uses a service from Adgorithms LLC, a marketing platform that uses artificial intelligence technology with its Albert software that aims to perform many of the manual, time-consuming tasks involved in a marketing campaign, from digital media buying to execution to optimization and analysis.

Connell says implementing high-tech marketing is more affordable than ever before. She says Cosabella implemented Emarsys, Sentient and Adgorithms all for less than what it was paying a digital agency before—an agency that didn’t use AI or machine learning in its marketing approaches.

“You used to have to hire custom programmers to build these types of technologies,” Connell says. “Now we are doing it for cheaper than for what we used to pay an outside agency.”

 

 

Favorite