AI is Opening the Door to Personalization at Scale
AI's impact and role at Coty, Inc.
Fred Gerantabee, Global Vice President of Digital Innovation, COTY
When we talk about AI's impact and role at Coty, the place to start is the end result for the consumer and, in turn, for the organization. There's a lot of emphasis now on brands becoming more like service providers and not just product companies. This is true for us as a beauty company and for the consumer packaged goods (CPG) category as a whole.
We can no longer just put items on the shelf; today's consumers want value and services heavily rooted in one-to-one personalization.
While bringing one-to-one personalization in communication aimed at consumers has historically been difficult to do at scale, we believe AI can help us achieve it in two meaningful ways:
Using data to drive product development cycle
Like many organizations, we've collected a ton of data over the last decade-plus, but we don't always use it, at least not as fast as we should. AI gives us more sophisticated ways to extract, classify and interpret not just hard test data, but visual data as well. As an example, we use computer vision, an AI field aimed at giving computers a visual understanding of the world, to analyze trends from the social media space. For a beauty company, extracting what's happening visually in fashion and culture is far more important and powerful than just pulling metadata or reading captions, where much of the context is lost. Using AI for insights and trend mining not only helps us be a more dynamic marketing organization, it guides our product research and development (R&D). As we try to shorten the product development cycle, we see AI as critical to determining what we should take to market.
Going beyond segmentation to create better consumer experiences
It's not enough anymore to assume that because Julia is 25 and lives in Chicago, Illinois, she likes the same things as 1000 other people with similar demographics. With AI, we can go further and aggregate first-, second- and third-party data to create highly dimensionalized profiles unique to each consumer. We can then have what feel like personalized conversations with each of them and use human-in-the-loop (HITL) machine learning to optimize and evolve those conversations over time to be even more relevant to each person.
The increased demand for value and service impacts everything we do at Coty, including
1) the creative and media we serve,
2) the conversations we have with consumers and
3) the data security we implement to protect our consumers and business.
In turn, we see AI affecting the way we approach each of these areas.
Data scientists derive insights for our creatives to build upon. We integrate these capabilities to get the most value from them.
Rethinking creative approaches and brand management
The “big idea” has always been core to the creative process for companies and agencies. While still important, what matters today is where the big idea comes from. It can't be rooted solely in one person's opinion or someone saying, “Hey, I've got a good idea;” it needs to solve a real problem or need. At Coty, for example, we're not just convincing consumers we have the right lipstick; we're answering a higher order mandate to make people feel good, confident and able to express themselves. With AI giving us a more dimensionalized view of the consumers we talk to, we think very differently about our creative approach. Instead of testing two or three concepts where we're either this or that, we marry the big idea with data to ensure it's meaningful to consumers. Functioning this way is actually much more effective for big ideas. Some of the best campaigns I've seen were based on insights from real data -- sometimes data sitting in plain sight.
This level of insights and intelligence leads us back to AI and the people involved in the chain. It used to be the creatives sat in one part of the building and data scientists sat in the other. That's a fail. They should sit in the same space and talk to each other every day. Data scientists derive insights for our creatives to build upon. We need to integrate these capabilities to get the most value from them. We must also think differently about doing things in line with the brand. We once saw brand identity as very superficial. We made sure our photography, fonts and creative were right and our message was consistent across media. It's much more visceral now. Our services, alliances, partnerships, consumer tools and platforms must all take into account what the brand stands for.
From a technology standpoint especially, everything must roll up to the brand. When my group thinks about a new piece of technology or a platform, we really have to root it to that identity. If a platform plays in a space that isn't true to the brand, or if we partner with a company that doesn’t support diversity the way we do as a company, it is not authentic to who we are and for our consumer. People know when brands don't deliver what they promise, whether it's a product attribute, personal service or a higher order mission.
What matters today is where the big idea comes from.
Evolving smarter conversations with consumers
Meaningful conversations with consumers are incredible sources of knowledge and insights. AI gives us the ability to optimize these conversations. At Coty, we've used a few straightforward, decision tree based messaging tools, but now we're using natural language processing (NLP) for speech-to-text translation and context. It's a great data mining tool because it helps us understand how people talk about our services and products and evolve the way we talk about them ourselves for a more personalized consumer experience.
Additionally, the launch of our Digital Accelerator has been another step towards transforming our digital capabilities across the organization. In 2018, we took this initiative a step further by bringing in external startups to the mix and making AI the priority area. We now look forward to working closely with the winners of the program -- Nudest, Skingenie and Glamtech, as well as other participant companies that joined us to build AI tools that will help drive more seamless consumer shopper experiences.
Putting data security in proper balance
Data security is second nature at Coty. We think constantly about where data is stored, how data is transported and breach contingencies. But keeping up with evolving vulnerabilities, including those made possible by AI, is a challenge. The types of data we capture, for example, are changing. The definition of personally identifiable information (PII) used to be straightforward, as in name, email address and Social Security number. Now, social identities and universally unique identifiers (UUID) are in the mix too. New devices and non-owned platforms also create ambiguity. Voice-powered devices introduce variables that test the limits of privacy. Many new data points, especially in chat platforms like Messenger, are being scanned. Who's the responsible party or owner of that data: the brand, the agency, whoever built the tool or a platform like Facebook?
Years ago, people were apprehensive about allowing devices or software in their lives. Today, despite all the security concerns, we are consciously or unconsciously willing to trade privacy for personalization, convenience and utility. Devices such as Amazon's Alexa and Google's virtual assistant can potentially listen to everything we say and know exactly when we order batteries and diapers. The question is how we balance the tradeoffs and mitigate as much risk as much as possible. AI can certainly help by accelerating threat detection and response.
The Future of AI at Coty
While AI's enormous potential is impossible to quantify, it will no doubt play an increasingly significant role at Coty. We've identified three areas of our business where we believe the technology's impact will be most profound:
Personalized and assisted commerce tools and services
We use AI to better understand what individual consumers care about and then tailor our services and experiences accordingly. We want to have conversations with consumers, wherever they take place that resonates and evolves over time based on their interests and needs.
Insights mining for communication and product development
Historically, R&D has driven many organizations but lacked access to critical data insights. We think the insights derived from AI are as pertinent to R&D as they are to the people creating advertising and marketing. We make it a priority to share this knowledge with our teams across the board.
Understanding and qualifying consumers
With AI, we can do a better job of engaging the right consumers with products that fit their aesthetics, lifestyles, ethical viewpoints, color choices and more. This is the first step to creating truly personalized experiences.
With all the upside AI offers, organizations must be cognizant of the challenges it can create internally and externally. If we create great consumer-facing tools but fail to optimize supply and demand, delivery or availability, our efforts will fall flat. From an enterprise standpoint, AI can help streamline these increasingly intertwined functions to ensure we meet each consumer's expectations and our organizational goals.
This article is originally a part of our guide, Powered by AI: Communications and Marketing in the Algorithm Age -- MSL and Publicis.Sapient's global study of 1800+ client-side marketing and communications leaders confirms that communicators are embracing Artificial Intelligence with a surprising level of confidence and optimism.
Fred Gerantabee is global vice president of Digital Innovation for Coty, one of the world's largest beauty companies. Coty's iconic portfolio of brands includes COVERGIRL and Rimmel in the consumer beauty category, Calvin Klein and Hugo Boss in prestige fragrances and skincare, and Wella Professionals and OPI for salon owners and professionals. With experience developing both technology applications and products for agencies and corporations, Fred has unique perspectives on AI's current and future roles at Coty and in business in general.