Three factors behind Cosabella’s successful AI deployment


Artificial intelligence is making significant inroads into retail. The use of cognitive technologies — a term that encompasses artificial intelligence as well as machine learning and related systems — shows positive results for customer retention and win-back rates. The number of positive mentions on social media has also been found to be higher for firms using cognitive technologies as part of their marketing programs.

High-end luxury lingerie brand Cosabella, with a bricks-and-mortar presence as well as a website, has always been an innovator. But the disruption caused by startups hasn’t gone unnoticed by the 35-year-old company.

CEO Guido Campello says many of the brand’s fledgling competitors “have a lot of technology and they’re starting to speak to the consumer in a rapidly changing way.” Getting in front of that disruption meant retooling Cosabella’s approach to marketing.

The right structure

After evaluating the marketplace and the brand’s long history with its customers, an internal evolution was the first step. “We took the marketing team and merged it with the e-commerce team, and our digital director took over marketing,” Campello says. The company’s IT function was also moved into the newly formed group, a shift that would position Cosabella to better capitalize on emerging technologies.

“It was a big step, because we went from a design-led marketing effort to a technology-led marketing effort,” he says. With technology front and center, the team next had to determine how they were going to crunch data, develop customer profiles and connect with their audience in a rapidly changing marketplace.

There were other factors the team also had to consider in their continuing transition toward a technology-driven approach. Cosabella partnered with an agency several years ago, only to discover it added an unwanted layer of separation between the brand and its customers.

Cosabella’s largely female marketing team was squarely in the company’s demographic, but they were handing off knowledge to an external group. “We said, ‘How do we bring that function in-house without hiring 200 people?’” Campello says. The company needed a way to crunch data into actionable pieces without significantly expanding the team or incurring a lot of overtime.

When the organization’s leadership looked at their options, they had a revelation. “We were going to be in a perfect position to be early adopters of AI,” he says.

The right tools

Cosabella created an internal system with several technologies, including an AI platform from Emarsys that was implemented in October 2016.

“Emarsys can leverage other programs to develop customer profiles, so we can understand who we’re going after,” Campello says. “Immediately we realized that things have become clear enough and user-friendly enough that even the non-technology-savvy people on the team could use it.”

Whether employees came from IT or the fashion world, they were soon adapting their language and way of working, shifting thought processes and workflows so they could leverage the AI learning that now forms a core pillar of Cosabella’s customer engagement strategy.

The shift to machine learning and other AI-related technologies has been a long time coming, and it holds the potential to fundamentally transform how retailers connect with customers.

“Over the last decade, most digital marketing automation technology has been basically predicated on a series of human-driven decisions,” says Allen Nance, CMO at Emarsys. The more traditional drag-and-drop approach, where marketers select a variety of consumer traits they want to target — gender, those who haven’t purchased in the last six months, etc. — brings with it an interesting problem.

“Retailers now have more data at their disposal than ever before,” Nance says. Lack of information isn’t an issue, but turning that data into actionable knowledge continues to vex brands. “We have moved beyond what we refer to as human-driven personalization, because it just doesn’t scale,” he says. Rather than relying on humans to select which boxes to drag into the marketing engine, AI can digest, distill and analyze terabytes of data.

Cosabella was eager to use the power of AI to forge stronger bonds with customers across various moments in their lives. How could the brand reengage with women as they moved along the spectrum, getting married, having children and developing different needs?

“It’s difficult to speak to each segment, with its different competition and drivers,” Campello says. Trying to create those varied customer profiles was a huge job; he says Emarsys helps the company “start understanding where those customers are coming from.”

The platform can analyze the data and determine which customer segments are likely to transition into other categories later, which customers have lost touch with the brand and where potential customers exist who haven’t yet engaged with Cosabella.

“We could never do all of that without using this kind of technology,” Campello says.

Cosabella was eager to use the power of AI to forge stronger bonds with customers across various moments in their lives. How could the brand reengage with women as they moved along the spectrum, getting married, having children and developing different needs?

The right implementation strategy

Knowing that non-technology people would be primarily responsible for using the platform day to day, Campello says it was important for Emarsys to gear the deployment to the team’s needs and knowledge levels. Ease of use and approaching the implementation using laymen’s terms were early priorities.

As his team met with potential vendors, “We challenged them to speak to us about how easy implementation could be.” Members of Cosabella’s team were also brought along for demonstrations and discussions “to see how confused they were and what their interactions were like,” Campello says. Unless those who would be using the technology could understand it, the deployment wasn’t likely to reach its full potential.

Given the important role the system would play in Cosabella’s efforts going forward, Campello knew the company didn’t want to be a pilot case for a vendor just entering the marketplace, as so often happens with emerging technologies.

“The concept from Emarsys is already a generation in,” he says. “They’re well-structured and funded, and they’ll teach us a lot.” Emarsys also brings a diverse team with global experience to the table, something that resonates with worldwide brands like Cosabella.

“They speak that global language already,” Campello says, “which makes us feel more comfortable in the markets we’re going after.”

Moving the needle

The list of potential uses for AI in retail is almost limitless, and the technology is still maturing. “It’s this idea of using artificial intelligence to bridge data science to execution,” Nance says.

Instead of simply giving marketers another report, AI can actually execute the campaign. The system will show customers a perfectly tailored offer or message based on their profile and what has resonated with other consumers like them.

“That is literally you loading offers and letting the machine decide which offer to show to a customer,” Nance says. Forget the human-driven guesses that have sculpted past marketing efforts; today’s AI is all about knowing.
Campello cites the numerous efficiencies gained with AI as a primary benefit to retailers. “You’re gaining a significant amount of intelligence,” he says. “Plus it’s doing work for us that our director was probably doing manually before.”

The Emarsys platform can compile and analyze data faster and with greater accuracy than a human, putting actionable information in the organization’s hands sooner. “Now the team can take it and react to it,” Campello says. Cosabella is still on a learning curve, but efficiencies are already appearing across the spectrum.

“It reduces brainstorming time and allows independent action,” he says. “We’re super-fast now.”

Julie Knudson is a freelance business writer who focuses on retail, hospitality and technology.


Comments are closed.