In September, Janet Sherlock wrapped up her seven-year tenure as the Chief Digital and Technology Officer at Ralph Lauren, the storied American apparel and lifestyle brand. Her position included full responsibility for the global e-commerce operation of the business. And she helped introduce AI into many of the business’ core functions, including product design.
“If you are a designer, you are going to sit there and say, ‘I want to build a sneaker, and I want it to have…’ That becomes your prompt,” she says. “What you would normally do is sit there with either a pen and paper or use some kind of Adobe tool, iterating, iterating, and iterating. Well, now you can do something, and you can save yourself 40 hours of time just getting to a concept.”
Sherlock joined Ralph Lauren with deep experience in retail and e-commerce technology, serving as CIO at Carter’s Baby Clothing and Guess, and working in custom home furnishings.
We spoke with her in July as part of our research report, Creating New Value in Large Organizations: What It Takes.
• • •
How does your department define value creation in 2024?
Running the digital business, I can tell you that we have very specific Key Performance Indicators (KPIs):
- Revenue is the bottom line;
- Operating income;
- Customer conversion is still a really big metric for us — how the consumers feel and whether or not the experience lends itself to actually purchasing something.
- Pricing. Your conversions are going to be lower if your pricing goes up; that is what is happening in our digital environment.
- New customer acquisition;
- Customer Lifetime Total Value (LTV).
We view churn differently in consumer retail than in enterprise software. We use churn to look at the customer file, and we look at time on site and length of visit.
We previously used average order value, but now we are looking more at the increase in order value. It is not really about the average but the incrementality of the increase. From my vantage point — a retail environment – customer conversion doesn’t count. When you are buying luxury, sometimes you go in with a destination, but, for the most part, people are shopping — whether in a brick-and-mortar location or a digital environment.
Does your organization have particular strengths or weaknesses in value creation?
Our website was acknowledged by Gartner’s Digital IQ Outlook for Luxury Retail as the top site for two years in a row. Our team does a great job, but we are losing a bit of revenue. We have global differences in how we represent our brands [and] I’d say that we have a lot of room for improvement.
In the world of AI, there is always more progress to be made. We call it a “beautiful tension.” We do have improvements to make in relation to brand representation and to our site’s overall performance, [although] we probably use more video than anybody else within our PDPs (product detail pages) and landing pages. We have improvements to make in our inventory optimization. I am leaving with so many things I accomplished, but there are so many things that are left undone.
How have you managed the response to the AI hype cycle, and what have you identified as the real business impact of AI as the voice of all things digital within a large organization?
We are actually ahead of the game because we had already invested in an AI/ML platform before ChatGPT was released. Our AI/ML platform works for us because it connects to all of the foundational large language models (LLMs): Anthropic, ChatGPT, Hugging Face, etc… we watch the inputs and outputs — and make sure that they are guarded, like no personal information and no hateful language.
We have our own data stores: one for store operations, a data store for legal that has all their contracts, one for procurement, [and] one for investor relations.
In the digital group, we have our own Ralph Lauren GPT. We haven’t turned off ChatGPT… We have our risks, but they are not as critical as in other industries.
The holy grails for fashion and retail are a couple of things: I would love to have a virtual try-on. But we have implemented it, and we are still going through our digital library of millions of images and describing them in great detail. It is only Chat GPT 4.0 as a tool to start, but you get great visual descriptions…
Our strategy has always been to democratize the utilization of AI. We have trained thousands of citizen data scientists.
What is a citizen data scientist? What is this army of data scientists you all created?
We say citizen data, data citizens, and data scientists. It is viewed as low-code and no-code development of models. What we do is train data scientists on our own model for, let’s say, product attribution and market basket analysis: What sells with what? We have a model, but then we want to train people on the model so that they understand how often belts sell with pants versus whatever. They can take the model and they can mold it to what they need. It gets interesting, because I’ve sat in on the training sessions for these things, and in the end, it is highly technical. You still have to have a decent amount of technical wherewithal. But the point is that you do not have to be a true data scientist. You do not have to know the Monte Carlo method, for example. You have to know just enough of the basics.
We talk about balancing magic and logic all the time.
Gone are the days when there were a few sacred people who were the knowledge base and held all the knowledge. You can no longer operate that way, nor do you have to hire people. It is a multifaceted way of trying to get to those “citizen data scientists,” and making sure that we are hiring people who have at least the gene for data — you still have to have the gene that says, I am data-driven. We talk about balancing magic and logic all the time.
Or take our designers. Our designers are amazing. I would’ve expected them to reject generative AI outright. But we have a couple of [colleagues] who just really took to designing with AI, and it is not really designed; it is more with the concept phase of the design process.
If you are a designer, you are going to sit there and say, “I want to build a sneaker, and I want it to have…” That becomes your prompt. What you would normally do is sit there with either a pen and paper or use some kind of Adobe tool, iterating, iterating, and iterating. Well, now you can do something, and you can save yourself 40 hours of time just getting to a concept.
It has been interesting. We could have received a lot of rejection of the AI-based tools from designers, but what ends up happening then is that you have a couple of people who suddenly are faster, better, and come up with better concepts, and they are taking those into concept meetings — and Ralph loved some of the concepts, so now all of a sudden everyone’s on board. It is huge.
How do you view the role of new technologies in value creation? Which new technologies are you looking at?
We haven’t put this into our consumer-facing website, but internally we have our own ChatGPT called RLGPT. It is a GPT interface connected to our data lake. Theoretically, beyond those data stores that I mentioned before- like store ops and contracts and investor relations – our data lake has all of our sales. Instead of pulling a report, you would say. “What were sales last week?” or “Why were sales down yesterday?” It can tell you yesterday’s sales trends; [that] we are seeing weakness in this particular market, we are seeing strengths over here. Then, we will add predictive analytics and maybe layer on the weather. It is going to be flipping amazing.
The most important thing to me is getting the technical and technological foundation right. Having your data in order can be an almost impossible task. But we had been working at that for years already before ChatGPT came out. You need to have technological architecture. It is really important to step back and look at it and say: How is the data being used? How are the models being developed? How are the models being stored?
At Ralph Lauren, the marketing team models are all built on the same AI/ML platform that every other model is built on… It is a technology strategy first. Then you layer on who’s using it, which comes back to the democratization of the technology. And then you ask, what are we trying to do with it?
Visual attribution is core to us. We scrape our competitors’ websites. We’ll look at it to compare what competitors are charging for, let’s say, a polo shirt. You can’t do that analysis unless there is a visual attribution that says, “What does a polo shirt look like”? We built a layer of technology for that capability. We built it for our websites to enable a complete competitive analysis based on visual attribution. It is a core technological capability that services so many different aspects of what we do as a luxury company. It is so important to have a technology and architecture model. So ask yourself: How are you going to do this data estate planning step first? Everything gets built on top of that.
Talent is at the core of everything. That is why we are democratizing AI. It is because we believe in the value of our employees.
What is the role of talent in value-creation efforts?
There are three things that make companies run: money, people, and technology. And no matter how you look at it, the most important of them is people. Without the people, money and tech would not happen… Talent is at the core of everything. That is why we are democratizing AI. It is because we believe in the value of our employees. We do not believe that we are going to buy it from somebody else and it is going to be better than what we know. We are better at taking our own data, our own knowledge, and our own direction and trying to apply the technology to it, as opposed to having people come in and tell us what to do.
What is one piece of advice you commonly share with other innovators in big companies?
Go with your gut. It is funny: I had a couple of startup ideas, and I didn’t have enough guts to pursue them. I think that both of them would have been really good. I look now at what other companies are doing and the direction we have taken with AI. I knew that this wasn’t the metaverse. There were many people when ChatGPT was launched who were asking me: “Is this hype?” I would tell them no, this is it. The difference is that the consumerization of generative AI has just happened. Generative AI has existed for decades. When you use Waze, that is generative AI.
So, follow your gut before the commercial phase of a technology hits… While you can listen to and track the hype cycles, you will need to take your own path toward it. You will see beyond what the hype is saying.
