Source: cosabella.com
Is AI the key to legacy brands’ revitalization?
Faced with a new selling landscape and the advent of direct-to-consumer apparel sales, family-owned luxury lingerie brand Cosabella turned to artificial intelligence (AI) tech. CEO and creative director Guido Campello told the story in a session at the Internet Retailer Conference and Expo 2018 (IRCE) in Chicago. He was joined by Vish Ganapathy, managing director at Accenture.
“We realized that we had to speak a different language, we had to use different terms, but there was no way to test everything,” Mr. Campello said.
With that in mind, Cosabella implemented technology like a machine learning solution, which allowed the brand to pursue advanced A/B testing, and an AI marketing platform to identify trends in how its promotions performed.
Mr. Campello said the company also used AI to help with email targeting and product personalization for its intimate apparel sales.
“The pattern recognitions were a big deal for us,” Mr. Campello said. “We would see these blips and be able to react. That kind of data crunching, we wouldn’t have been able to accomplish [without AI].”
AI allowed the company to hone its images and ad copy by recognizing trends, A/B testing and acting on them the same day. Mr. Campello showed examples of the website, with its AI-optimized look, feel and messaging.
As an example, Mr. Campello said that a 35,000-variable test had revealed that customers received the copy “family-owned since 1983” far better than other options, such as deal offers.
“Everything you see on our site right now are things that I would have never approved of,” Mr. Campello said. “I would have looked at aesthetics differently, I would have looked at script differently, I probably would have looked at language differently.”
“AI itself is simply a means, it’s not the end,” added Mr. Ganapathy. “It opens up the aperture in what else you can do. It opens up the possibility for you to analyze data in a much wider manner.”
For a brand like Cosabella, which was dependent in its heyday on celebrity models like Sharon Stone and Heather Locklear, this ability has opened the door to successes like the rollout of an entirely new product category for an AI-identified segment.
Discussion Questions
DISCUSSION QUESTIONS: How useful do you see AI technology being in turning around once-popular brands and retailers? What opportunities and dangers do you see in using AI to formulate marketing and branding approaches?
Poll
BrainTrust
Ryan Mathews
Founder, CEO, Black Monk Consulting
Laura Davis
Founder, Branded Ground
AI is a great, relatively new tool and it does enhance a retailer’s ability to refine its marketing campaigns, advertising and its website, maybe even its pricing. let’s not forget, however, that it all starts with desirable product at a reasonable price.
The key here is well stated by Mr. Ganapathy, “AI itself is simply a means, it’s not the end.” Artificial intelligence technology increases your decision options, greatly accelerates your ability to (re)test and apply, and challenges current models of thinking while infusing new agility into your organization.
Is AI a revolutionary tool and lever? Absolutely yes. The turnaround silver bullet of retail? No way.
Hats off to Cosabella for sharing their success in generating insights with AI-enabled consumer research. This is the way of the future. As it grows, brands and retailers (or their agencies) will develop the skills to create the structure for information gathering and, more importantly, the ability to graphically represent insights gathered. The rewards are fewer wasted marketing dollars, less de-branding, better segmentation and better sales success through improved relationships with targeted consumers.
Without knowing for sure, I expect Accenture helped them take a step back and think about business (or marketing) problems that lend themselves well to new forms of analytics, like AI. Then plotted out a use case path to test, rollout and scale.
When companies drive right to the technology “thing” as the solution to their problems, they typically fail and AI analytics is no different. Longer term, companies of all sizes and maturities have opportunities to apply AI to many elements of their business to improve results. It’s what really large companies like Walmart, Starbucks, P&G and others have landed on after learning while they go over several years.
Companies that are less progressive can learn from their experiences to accelerate the value and scale they achieve with AI — in marketing, supply chain and many other elements of the business.
I like the idea of same-day reactions to A/B testing, and congratulate Mr. Campello for being brave enough to use the data to make content and marketing decisions that he would have otherwise vetoed.
I just read an article in WWD yesterday that said that 86 percent of all AI home use was still for “daily tasks,” like search, rather than for actually purchasing something. So regardless of the “new shiny toy” syndrome, it’s going to be a while even for marketing. The only saving grace was that Millennials and Generation Zers were tops with AI use for purchasing. But ultimately, it’s going to be a lot slower of a transition than we all might’ve been excited about initially.
The decisions made by Cosabella seem much more to be data-driven ones by a company that was committed to change. It’s hard for me to call this kind of segmentation and presentation AI, but I may be missing a nuance somewhere. Sounds like pretty straightforward data analysis.
Most importantly, congratulations for having the determination to change!
AI is a tool and tools can make work easier or they can be misused. So depending on the application AI has a high potential to help turn around the fortunes of once popular brands. But that requires developing good internal talent and/or the right vendor/partner selection, aligning systems across the organization and — of course — getting the software programming right. I think we are really in the infancy of what AI could be doing in terms of marketing and branding which means two things are probably true: in some respects we are overestimating the power of the tool to autonomously solve all of our problems; and, we are also unaware of the real power AI brings to retail problem solving.
Over time the guardrails for employing AI and the true dimensions of its potential and limitations will become clear.
Funny to see this product as the feature, as they have definitely gotten my attention in the past months. As shared, “AI itself is simply a means, it’s not the end.” They did all the right things around their AI test. The models became less Barbie-inspired; the copy became more human; the product name is more culturally relevant rather than cult-like; the price point became more justifiable.
AI is a very, very strong tool to get the right product closer to the right people. But once the hook is in, the rest of the experience needs to pay off again the holistic CX plan — which includes channel experience, voice/tone, brand feeling, product design, product story, price, transaction ease and all that jazz. No matter how, when and why it’s used, AI needs to be a lever to aid in creating a more holistically human experience.
AI is allowing Cosabella to make objective decisions about their products and how to market them. Other retailers could develop similar analyses to more correctly predict where they should position their products and marketing messages. The real challenge is to be able to gather the right information for the models and interpreting the results properly. If the data collected and the models built are not properly done the results will be skewed from reality. The models should be tested in pilot programs to ensure they reflect reality before implementing them across the board.
The term “AI” is suddenly applied to all things tech. One has to dig deep past the rhetoric to understand if a new technology is really segmentation in disguise. That being said, I think Cosabella is way ahead of the curve in re-branding itself for a new generation. An open mind, embracing both human insights and technology to create a view of a brand’s customer preferences is a solid day of work at the office!
It is great that legacy brands like Cosabella are using data mining tools to identify trends and assess the success of different marketing initiatives that I am sure would pay dividends for many similar brands heading directly to the consumer. However, this is not AI. It may be sold as such because it enables consultants and software companies to charge more but it is not AI. Machine learning is about understanding what is happening and automatically making changes to the activities and decisions made without having to rely on human intervention. What has been described here is data mining unless there is much more to it than explained here.
I am not making this point to be smart, but because there is a huge amount of confusion around AI and many companies are stating that they will not take a particular solution unless it has AI without understanding what they actually mean. As said, it is a way of upping the price, the consultancy rates on projects, and for many software companies how to inflate your PE.
Please let’s get some fact and quality discussion on AI which is immensely powerful as used in supply chain optimization and other areas, but let’s not confuse it with data mining.
Was wondering the same thing Andrew. Seemed to me like simple pattern recognition and decision trees. Maybe the biggest feature of AI is a rebranding of “data discovery/ data mining” to something cooler.
In retailing, AI is already being employed to drive brand value for new and old brands alike. However, we need to remember that AI in itself is not the tool. AI augments the intelligence of software tools. First you need to identify what you aim to accomplish and then understand what information you need to capture. From there, using AI capabilities, like natural language processing, visual recognition, etc., it can help users make decisions around the shopper experience, digital marketing and myriad other functions. I don’t see any real dangers in using AI to augment our decision process. I only see more and more opportunity as humans better understand ways to improve retailing effectiveness overall.
I’ve been on a national tour to engage retailers in talking about AI and its ability to improve their performance, and my sponsors and I are finding that companies have a hard time articulating use cases for this technology. Thumbs up to Cosabella for sharing their success and inspiring others to push ahead.
This sounds like basic analysis. What am I missing?
This is a tool not a solution. It is another tool to better understand consumers. If the algorithms are good, if the analysis is sound, and if the insights ring true with consumers, it is another extremely useful tool.
I wonder how much is down to the “AI” and how much is down to Cosabella just actually making use of data that most retailers have but never do anything with — or not in a very speedy fashion. I think their use of machine learning to quickly react to different customers and scenarios (on the same day!) is a real smart approach to all of the data that they’re gathering.
I really think that the key is how quickly they’re able to action changes based on the information and recommendation from the AI — if they were taking weeks to tweak things each time I’m not sure you’d see the same results. I certainly wouldn’t say that AI is going to fix all legacy brands’ problems, but if they use the tech right they have a huge amount of data to mine and benefit from.