Using data to predict fashion trends

Publish date:

Data in vogue. Could a data scientist be a fashion pundit?

Reatilers have always looked to fashion designers, celebrties, etc. to predict new trends in fashion. Could this be done through data?

This edition from our New York Applied Innovation Exchange features Karen Moon, CEO and Co-Founder of Trendalytics, a product intelligence platform that allows retailers to understand and predict trends by aggregating billions of consumer demand signals.
We were first introduced to Trendalytics during a Retail Pitch event hosted at the AIE and continued to collaborate during the National Retail Federation Big Show event in Jan 2019 and now for more specific client engagements.

In the conversation with Karen, Peter Maloof, Director of the New York AIE, dives into Trendalytics, the power of data, and how social media platforms are altering marketing, advertising and merchandising in retail.

Also available on:

 

 

 

Transcript:

Peter Maloof:       

We’re kicking off the latest installment of Capgemini’s Applied Innovation Exchange podcast. And this afternoon we’re sitting with Karen Moon, who’s the CEO and co-founder of Trendalytics. Welcome, Karen.

Karen Moon:         

Thanks for having me.

Peter Maloof:       

Thanks for showing up and helping us out here. Want to just talk a little bit about you and Trendalytics and where you see the industry that you’re working with going. And we’ll just jump into it. Give us a little bit about your background and a little bit about Trendalytics.

Karen Moon:         

Sure. Really quickly, what the vision of Trendalytics is to build a predictive analytics platform for consumer products businesses. And we work with clients, a number of Fortune 500 companies, with their merchandising and product development teams, to help them make better, more informed decisions. But what’s missing was understanding where consumers are at. I think often times in the business, and this kind of leads to my background and why we ended up here, but we wanted to take a different point of view with big data. Because what I saw happening is, in a lot of the organizations where I grew up in my career and lived, oftentimes you can, the challenge is you can be too analytical and you forget the story behind the numbers and truly the art and science of the business.

And so if you think about a lot of the companies we work with, in apparel or fashion and beauty segments, it’s constantly changing. But these are organizations where it’s very creative led. How do you convince a designer or merchant that they should understand and listen to the pulse of the consumer? That’s something that’s totally changed in the last five years.

Peter Maloof:       

Right. You’ve got kind of a sexy aspect of big data, which is, you’re using it to apply to fashion and also you’re extending beyond kind of fashion. Hence the name, Trendalytics. You’re trying to stay on top of that. Your company provides something that I think it’s a little different in terms of being able to look a little bit more predictively at things rather than using just point of sale data or kind of lagging indicators. How did that come about? And how did you get that built into your platform?

Karen Moon:         

Yeah. And then if I take it all the way, a step back to your first question, because I don’t mean to avoid that, what am I doing here? Like how does this all fit to me? What I do. Well, actually my first job ever, because you asked me a lot, as a kid and I always wanted to be a fashion designer. I always like to draw. I used to make clothes for my Barbie dolls. The first thing that was the easiest thing to make with scrunchies.

Peter Maloof:       

Scrunchies for your Barbies.

Karen Moon:         

Which are cool, but again now. I made scrunchies and I was always kind of like, I don’t know, I just always had an interest in making things and commerce. That sounds cool now. But then that was my first real kind of fake job. And then I actually went to art school, but then I dropped out.

Peter Maloof:       

Okay. Where did you go? Where did you start art school?

Karen Moon:         

It all comes together, I promise. I went to art school, I went to Otis, which is in Los Angeles. It’s a Fine Arts school. They have a fashion program. I sucked at it. It was really hard. I was there for a year. I did okay. But it was, it was really hard. I think art school is harder than investment banking in many ways because you can never, it’s like coding, you can spend a lot on that project and you never know when you’re done. But it was really challenging in different ways and it made me look at the world differently.

It made me things about composition, photography and really understand the creative process. And so I think that was an important piece of my journey because I did that for a year, dropped out, went to regular school and then I started my career at Goldman Sachs as a technology investment banker. Now it all makes sense.

But what’s different about the way we approach data at Trendalytics is, if you think there’s always creative people in the organization and there are the more, there’s left brain and right brain. But the power of it is when you bring left and right brain together and there’s very few different types of people who can understand that balance.

And so as we think about the business of what we do in consumer trends and anticipating those, there really needs to be the balance of understanding how that art and science comes together so we can give numbers and data and graphs. But really when you think about consumer trends, there’s this element of how and when do you bring in newness and how do you make it specific to a brand?

If every, if Chanel, Gucci, Abercrombie and Fitch and Kohl’s all made the same widgets, it wouldn’t work. There’s this element that they make it right for their brand and their consumer set. And so how do you enable that and how do you empower that? In a lot of ways what I saw the opportunity in data was that there’s a lot of companies where they were approaching analytics from a very black and white perspective. But there’s these other fundamental gray areas and there’s this element of actually how do you provide quantitative insights and confidence?

For one, you should be sunsetting your bestsellers and actually empowering more risk taking ideas. Because that’s really what’s moving the business forward. Whether it’s in consumer product industries, whether it’s in fashion and food. Understanding the sea change or an inflection points when a consumer preference is changing is what makes the difference between the brands that are continuing to grow and continuously relevant and the laggards that are kind of falling behind.

Peter Maloof:       

Okay. And you mentioned a lot of really large brands that have, obviously very large paid presence in places. And yet I think some of the power of what I’ve seen that you’ve got is that you guys look at influencers and other things. We all know that brands with purpose and smaller brands are really meaningful for this generational cohort in what they’re looking for. How do you kind of work with that aspect of it?

Karen Moon:         

Yeah, so it’s interesting. When we, the reason we look at some of the data sources you do is that we’re looking for leading indicators of demand. What are the information sources that we look at? Well first of all, what is the problem we’re trying to solve? We’re trying to say, if you are in the cosmetics space, then what are the products that are trending in the marketplace? And what are the products that you should make? But also what are the skincare concerns? Are there certain treatments? We really think about the ecosystem of the consumer and what are they searching for? Whether it’s in a mascara or eyelash extensions. Basically, what are the substitutes and the compliments around that? Are they applying their makeup differently? And where are they getting inspired?

And I think this is a really good example of a place where for a long time people didn’t know if social media was driving sales. For a long time they didn’t know if influencers were relevant. But the thing ism we’ve always had influencers, whether they were celebrities, singers, magazines. It’s always been a fabric of commerce. And so what we look at at Trendalytics is, well, who are the ideas, the media companies, the celebrities, the bloggers that are inspiring newness? And so we think about what are interesting moments in that consumer journey? Where are they inspired on social media? Who are they following? What are they talking about? And are they talking about scrunchies in a more interesting way? Is it on Stranger Things?

And so then more and more people are talking about scrunchies. Are people searching for scrunchies? Is it accelerating? How many scrunchies are in the marketplace? At what price points? Are they a velvet, are they whatever? And is it in the luxury segment? Is it stocking out? Is it in Nordstrom? And so all these things together, you get a really interesting sense of the consumer adoption curve. And so then you look at scrunchies and you say, well what about scrunchies relative to the banana clips? I’m using all these 80 references.

Peter Maloof:       

That’s awesome. Every single you’re saying.

Karen Moon:         

And so the whole idea is, it’s almost like managing a hedge fund portfolio. I have $40 million to spend across which companies, and if there’s competitor companies, which one am I placing my bet on? If there’s some that are complimentary, then how am I going to understand that. And so it’s really about understanding these ecosystem of consumer purchase patterns, behaviors around products, and then driving that. And so now you can come back to brands. What’s interesting is yes, I can look at all these different types of hair clips or whatever, but then you also want to know sometimes brands actually start those trends and they create newness. And I think the direct to consumer brands that have, are educating the consumers in different ways now. I think generationally we’re going to see a whole host, we’re already seeing all, every consumer product category is being reinvented.

But what it’s doing is it’s educating the consumer around that new segment. And if you look at Casper, what I think is so interesting is, they are basically teaching a new generation the importance of sleep. And it goes back to one of my earlier jobs, my other random job that this was a real job was I worked at, guess what department I worked at in Nordstrom’s when I was in high school? Men shoes.

Peter Maloof:       

Of course.

Karen Moon:         

And so one of the things I learned there was well, the reason why is they’re actually genius. When I was in high school, I was on the Fashion Board. Basically all the Nordstrom’s school.

Peter Maloof:       

Hold it, what high school has a fashion board?

Karen Moon:         

All the Nordstrom’s recruited high school students. But they were literally part of their focus groups. But I wanted to work in Nordstrom’s, dah, dah, dah, dah, dah. And so then, and then you also get a job there. And so in the summers you get to work there. Then I was a cashier in the men’s shoe department and I saw how much money everyone made. And if I asked the guys, I was like, “Hey, can I be a salesperson next summer?” I was selling men’s shoes. I think everyone thought I was cute because I was so little and I don’t know. And I was selling all these men I don’t know, Allen Edmonds and stuff.

What I learned is you should spend a lot of, you should invest in shoes and a mattress because you spend half your time sleeping. Importance of sleep. Anyways, so at Casper, they’re teaching that there’s a whole new, we actually coined a term at Trendalytics, “Sleep is the new wellness” in 2016. Arianna Huffington, everyone, mindfulness, what they’re also saying is the quality, these are things that we should invest in. You should replace your mattress every so years. The importance of sleep and bringing the sleep lab. It almost took a commodity product and created this whole experience around why it’s so important. And I think the same thing is happening across all these other categories. I think the most boring categories are the most interesting, like the electronic toothpaste or toothbrush company.

Peter Maloof:       

Toothbrushes.

Karen Moon:         

Are super interesting. It’s you should brush your teeth in a different way. And it’s just there’s all these consumer products like mints that are being reinvented.

Peter Maloof:       

Right. There is this direct consumer model. The barrier to actually have a global distribution, certainly a nationwide distribution, it’s so easy right now. Whether you’re going to go through Amazon web services, inhabit that way or whether you set up your own Shopify or anything. You can see it’s easy for people to be able to do that. And so what I hear you saying is that that direct to consumer then influences the other manufacturers and they can start seeing what those trends are. And you do things that are adjacent to fashion such as, well, the Arianna Huffington thing of talking about sleep then persisted into all these other areas. Then other influencers were then talking about sleep and then that kind of led to Casper and Purple and all of these direct to consumer brands blowing up.

Karen Moon:         

What’s interesting is if you take that macro trends, so what we’re doing is obviously we look at micro trends but we also see how they roll up culturally because that it’s just a fun part of what we do.

But it is interesting to see all those patterns. We looked at, we always like to look at the new brands because part of the trends for our retail clients is, well we have retail clients and just brand clients and for our brands, they want to know who their competitors are. Who are the upstarts that they should know about? For our larger retail clients, they want to know for the up and coming brands. Having a product trend is just as important as having a brand trend because sometimes brands are cool and a certain brand with a cool t-shirt can do better than, I won’t name names. Other brands. It matters because there’s an emotional connection. And so what’s interesting is when you think about some of these macro trends, even using sleep, there was this whole education around sleep, peak performance, and you think, all these other things in the ecosystem that drive around that.

And actually that is something that actually trickled into fashion. We talked about that as a cultural trend that even back then we you started to look at pajamas and how that was seen, early indications of that happened in fashion and it’s still happening three years later. The other one was he saw slip-ons where Gucci, which is a very influential brand and drew a lot of stuff, had those slip-ons. And there are all these elements around comfort, cozy, cliquey. There’s all these things happening together. And so you look at the micro trends and how they roll up with macro and how that gives a symbiotic relationship.

Peter Maloof:       

Right. And that’s good. That’s really, again, more of the art and science. You started with kind of the foundational science and looking in a particular area. And then saw just things that were happening at Jason and knew that they would actually influence the data that you were looking at.

Karen Moon:         

And so some of the other stuff we do on the trends, we just like to look at consumer behavior and it’s just a fun thing we do on the side. But it’s interesting as we start to do, we’ll do two reports a year on how do we think consumer behavior trends are changing? What that means. An interesting customer co-creation, we’ve been talking about that for years. What does that mean in terms of newness and what consumers are expecting? And so it’s really more about meta trends that relate to how can, what is the new paradigm of how they’re shopping? What are the new business models that are reshaping how they think? And at the end of the day, that does impact what the larger brands and retailers have to think about from an experiential standpoint in their stores or online. And so it all does kind of tie together.

Peter Maloof:       

How often do you, you say you publish it twice a year, when do you publish it? And where can people find it?

Karen Moon:         

Yeah, so we do monthly intelligence reports and that’s where we’ll do something like this or what’s happening in special sizes or what’s the update on street wear? Stuff like that. But the consumer trend report we’ll do, well we actually do it once a year. We just published it a couple of months ago. Can send it to you guys. But for our clients, we actually did it mid year so they could think about what it means for the holiday. This was client only. We’re like, here’s what we’re seeing that’s cool. We’ll look at our data. But we’ll also look at what’s happening in Kickstarter. We’re always like, what’s not even in the market yet? And what are brands that are doing cool things? And what are new ideas that you can do from a holiday planning standpoint? We kind of figured out how do we leave it in to give our clients an edge?

Peter Maloof:       

That’s awesome.

Karen Moon:         

And then we’ll do it more publicly. And it’s just fun.

Peter Maloof:       

Yeah. I’m definitely looking forward to checking that out. Let’s fast forward and think about you and Trendalytics two, three years from now, what are your goals? What do you hope to happen?

Karen Moon:         

Yeah, so it’s interesting. When we started the company we definitely thought, well there’s a real need in merchandising. As I saw the marketplace, there’s we say 90% of the world’s data was created in the span of two years. IBM, that was what, 2015? That number grows 10x every year. Now it’s all the world’s data. But everyone was really focused on advertising and marketing. And so although the way we acquire our customers has changed and our organizations have shifted internally in terms of the digital transformation, it was all in that area. But we didn’t focus on merchandising and we didn’t focus at the core of why consumers are coming and buying our products. I thought that was the big opportunity. That’s the next way. We were way too early for the market. Now people are coming onboard. And so the timing is good.

But what I didn’t realize that we had to solve for is this data structuring issue. And so this was a big aha moment in the last couple of years where we are very fortunate to work with some really cool clients that are building our product with us because they get our vision and are really early adopters. And so it is this partnership. But when they told us they wanted to integrate our insights with their data, what we realized is how crappy their data was. How unstructured it was. And so now, and in 2019, we’ve gone more public with it. Where we’re saying, the big problem with retailers isn’t just that they don’t understand their client, customers and what they want and it’s hard to predict them, they can’t even see that pattern in their own data because it’s not structured that well. One of the aha moments for us is, we’re looking at a whole new lens of trends where it’s consumer led.

Not just what is next in fashion, what’s next in vegetables? Is it going to be broccoli or cauliflower? Like what’s the next thing? It’s more of when is it going to drive commercial? What is the commercial opportunity? When is there magnitude? And what is the relativity of that? As we look at that, what we realize is, we’re really good at looking at consumer trends, what people are calling on social media, what consumers are searching for. But that is very disconnected from the way retailers are describing their own products. One of the biggest problems we had to solve from a machine learning perspective is connecting the dots. Understanding how consumers are shopping for things, searching for things, how they’re describing products and the products that exist in the market.

As an example, I love using mom jeans because the biggest trends in data and men’s shoes have been mom jeans, boyfriend jeans, dad shoes, dad hat. It’s girlfriend. It’s just it’s all about the family. But these are the highest searched terms. Right now mom jeans is ranked number one in terms of the number of people searching for on a weekly basis, but no one is calling it mom jeans. Here’s the thing, you have your merchants, you have your fashion people saying, “The mom jean’s hot.” Actually the publisher of Vogue, I’m friends with, we met at several years ago, she was like, it’s all about the mom jeans. But it’s about, that was four years ago. She was right. But then if the editors are calling them mom jeans, if the bloggers are calling them mom jeans, I think Eva Chen, Instagram fan, she talks about mom jeans almost every week. That’s happening.

What if you can’t see if your mom jeans are selling out? If you don’t appropriately structure your data that way. What if you are top social media posts, all your marketing dollars are going to mom jeans, but you can’t analyze your sales? This is the disconnect. And so that’s the biggest opportunity that we’re doing at Trendalytics. Where does this take us? I go in tangents.

Where does this mean for us five years from now? What we’ve done is we’ve gone very deep in the vertical. We are looking in apparel, in denim, we’re looking in skincare and we’re saying, what are all the ways that consumers shop for these products? How do we really understand demand? And understand on a more intimate level, at the deeper level and across intent? Which is how they’re shopping for it and the products that exist in the market. This whole idea of relations is actually taxonomies and ontology. That’s a very hard problem to solve. That’s really what we do around our market verticals. But I think the opportunity in the future is the technology we built is actually going to build dynamic ontologies or taxonomies and understand relationships between product attributes in a given vertical. But we can do that in multiple verticals. We can do that in multiple languages. I think the growth opportunity for us is actually in the future taking what we’ve done and then applying it in new vertical markets.

Peter Maloof:       

Do you also do visual recognition? Because somebody who’s calling something mom jeans, perhaps that’s a high waisted jean. Maybe the vendor’s just calling it a high waisted jean. How do you know what’s a mom jean versus high waisted? Or maybe they’re the same thing and then how does that translate into the taxonomy or the ontology?

Karen Moon:         

Yeah, that’s a really good question. What we do is we actually understand if things are related or not. We look at all these different key words and images. We, to answer your question, we do a mix of image and text. There’s more we have to do in each of the areas. Some of it is, some of it’s today working and then there’s just an element of where it can go further. And so what we realized, to answer the question on a broader level, there are certain thing that text will always do better than image. And there are certain things that you can never describe with text. And so you need image. But then there’s also challenges in image recognition that, so the combination of the two is really helpful. Color is really much easier with image, as an example and then certain frames depending on the use case you’re trying to do, visually similar images is actually not a hard problem. But, and it’s subjective.

Sometimes people gravitate towards color. But there’s all these other things we’ve done in our image testing where the pose, the skin tone, all these things impact it when you’re doing object detection. And so it’s interesting but when you start to train your data set on a very specific area and so I think that’s the value of Trendalytics and all these other vertically focused platforms, is you can use the latest cutting edge machine learning algorithms because once you create those, people are creating these things, but it’s all in the data sets that you train and all in how well you do that and verify that. And so that’s all the grunt work of what we do in our vertical markets. I really think the future of machine learning and AI is solving very specific application problems. Not for the sake of using a new model.

It’s what are the problems we’re trying to solve because if I was an image recognition company, I would only focus on that. But because we’re focused on, what is the key problem we’re trying to solve for our clients? And we have very specific questions, then it forces us to say, what is the smartest way to answer that question? What are the best models to use for each funnel of that pipeline? And then it gives us a little bit more flexibility and freedom. But there’s also a lot of work to do. And so we have to stage our product development and technology pipeline to figure out where are the easy ones and how do we stage things?

That’s why we actually have never talked about ourselves as being predictive. And even this is early, but we are getting there. And so we’ve been good at things but now we’re actually testing our confidence intervals and all this stuff. And so all this work we’ve been working on for years, we’re finally at a place where we’re starting to talk about that in a very different light. We’re, it’s been under the radar, but we’re really excited about that.

Peter Maloof:       

All right, well you’ll have to come back and talk to us a little bit more when you get that predictive side down. I’d like to wrap this up and just kind of turn it back to you personally. If you weren’t doing this, I know that you had kind of a, you grew up on the west coast, though you’re here now, what would you be doing? If you were to sell this off or if you were to hit the lottery, where would you go? What would you do?

Karen Moon:         

Honestly, I think this is really fun. I like it. We come up with ideas all the time. I’d want to start a fund that helps people do that with data. And so incubating brands and entrepreneurs is something I love to do.

Peter Maloof:       

That’s awesome. Cool.

Karen Moon:         

And then connecting them with all the right people and the buyers, because we’ve created now, our own network of all the manufacturers, we’re friends with all the buyers and stuff. And so I feel like we can really help accelerate brands. And I used to be in private equity and I miss investing.

Peter Maloof:       

There you go. Well thanks very much for spending a little time with us, Karen Moon from Trendalytics, really appreciate it.

Karen Moon:         

Thank you for having me.

cookies.

By continuing to navigate on this website, you accept the use of cookies.

For more information and to change the setting of cookies on your computer, please read our Privacy Policy.

Close

Close cookie information