Subscribe
Share
Share
Embed
On the CosmoFactory podcast, discover the latest innovations along the cosmetics and personal care supply chain. Hear thought-provoking conversations with top beauty industry experts from around the world. Learn about next-level solutions and find inspiration to turn your own ideas into industry-changing innovations.
A PRODUCTION OF Cosmoprof Worldwide Bologna
CosmoFactory is the first podcast from Cosmoprof Worldwide Bologna—the most important beauty trade show in the world. Dedicated to all sectors of the industry, Cosmoprof Worldwide Bologna welcomes over 250,000 visitors from 150 countries and regions and nearly 3,000 exhibitors to Bologna, Italy, each year. It’s where our diverse and international industry comes together to build business relationships and to discover the best brands and newest innovations across consumer beauty, professional beauty, and the entire supply chain. The trade show includes a robust program of exclusive educational content, featuring executives and key opinion leaders from every sector of the cosmetics, fragrance, and personal care industry. Cosmoprof Worldwide Bologna is the most important event of the Cosmoprof international network, with exhibitions in Asia (Hong Kong), the US (Las Vegas and Miami), India (Mumbai) and Thailand (Bangkok). Thanks to its global exhibitions Cosmoprof connects a community of more than 500,000 beauty stakeholders and 10,000 companies from 190 countries and regions. Learn more today at Cosmoprof.com
CosmoFactory was co-developed in collaboration with supply-side expert Deanna Utroske, Host of the CosmoFactory podcast and Editor of the Beauty Insights newsletter.
Deanna: [00:00:00] This episode is about acne care. It's about digital skin diagnosis and the classification and treatment of pimples. It's about adaptive skincare routines, about access to information, and about the implications of consumer beauty tech, not only for omni-channel retail, but in the clinical testing sector as well.
Today, in the CosmoFactory recording booth at Cosmopack in Bologna, Italy, I am speaking with Elena Setero, co-founder and CEO of Dermaself. [00:01:00] Elena, welcome to CosmoFactory.
Elena: Thank you. Thank you for having me.
Deanna: Yeah, I'm so glad you're here. As I understand it- Uh, Dermaself is a configurable tech tool for acne care.
You essentially partner with brands and customize the platform accordingly. But Elena, you built it. I- how do you describe Dermaself?
Elena: We can describe it as a portable dermatologist that comes with you whenever you go beauty shopping, right? So it's like having a dermatologist whispering, uh, into your ear which products are the best for you whenever you are shopping, you know, at Sephora or whatever.
Deanna: Mm-hmm. And can you tell us a little bit more just about the company itself? Where are you based? Um, when did you get started? These sorts of details that will help us think about it.
Elena: Absolutely. So we are a startup, and we are based in Milan, Italy, as you can tell from my accent, I'm sure. Uh, we started out actually w- as a B2C personalized product, let's [00:02:00] say, uh, back in the days, and then, you know, the world changed.
Everyone, uh, understood really the potential of AI, and we understood as well that AI wasn't an accessory for us, but it was our main asset. Mm-hmm. So we pivoted to a B2B SaaS, a software as a service company.
Deanna: Perfect. That's very helpful to understand. And, uh, so many founder stories are, are very problem solution stories, and I, I don't always ask founders, um, for their background because they're...
To, you know, to be honest, there just is so much repetition in the, in the storytelling. Um, and, and, and we like the real innovative stuff here on CosmoFactory. Um, but yours is a bit more of a story about having a solution and then finding the way to scale it or bring that solution, really share it with other consumers.
Just briefly, will you tell us what inspired Dermaself?
Elena: Absolutely. So in my family, uh, very severe acne runs in the blood, and my mother became a dermatologist in order to treat her own skin. So myself, I h- [00:03:00] I do have a very severe acne too. Uh, but because I have a dermatologist at home, I'm very lucky, so I was always able to choose the right product, the right care, and always deal with my, um, acne breakouts, let's say.
Uh, and then once I was in London, because I moved, I had moved there for my work, and, you know, there was COVID, and there was Brexit, and really I didn't have a- access to my mother's products or prescriptions, you know, and so on. And there I found out what it means not to have a dermatologist with you- Yes, yes
every day. And, you know, my skin just got very angry, and I wondered how do people that do not have a dermatologist mother act? You know, how do they deal with this severe acne if they have it? And I answered myself that they, they can't. You know? So I told my mother, "I'm going to come back to Italy and we're going to build something for these people."
Deanna: Yeah. Yeah. [00:04:00] No, that's, that, that's wonderful. It's, it's really an instance of technology providing consumers with access to information, right? In your case, it was your mother, a literal sort of person in the next room perhaps. But that same information can be shared through technology. You know, in preparing for our interview, I, I remembered, uh, you know, hearing conversations maybe in the 1990s about how digital technology would help bring the best of medical care to rural communities, uh, as an example.
And I think your platform is doing something along those lines, reaching people where they are, and then advising on care or treatment routines that are available to them. Uh, does this sound right? And, and can you talk a little bit more about the current reach and, and potential of the platform?
Elena: Absolutely. So as I told you, uh, it all started with acne. So, uh, you know, uh, we started from the assumption, the dermatological assumption, that each pimple is different. So basically, whenever we're talking about acne or a pimple, we're [00:05:00] actually talking about at least eight different things, because you have comedones and papules and so on.
And each type of pimple requires different active ingredients. So whenever you go to a pharmacy and you ask for a pimple cream, an anti-acne cream, and they give you just, you know, a generic one, and they don't even ask you or look at you to understand your skin, uh, th- that's a big mistake right there.
Because, you know, you can use a salicylic acid if you have comedones, but you can't use it if you have papules, for instance. So we are using a lot of aggressive cosmetics on our skin, right? Especially after COVID. We all started doing, you know, 12 steps skincare routines. Like, the average woman uses 17 cosmetic products a day.
And we combine them without an actual logic because none of us can really read an ingredient list, and that's when Dermaself comes in, right? Mm-hmm. So
Deanna: we- I, I wanna jump in, because you've mentioned these, I think you said eight different types of pimples, [00:06:00] and I think this is something particular about your technology that might be helpful for us to understand.
Can you talk about sort of classifying those, how AI actually understands them, maybe how that compares to what we think of as, um Other, you know, skin analysis engines in the market.
Elena: So we do really know skin. You know, our AI was trained by cosmetologists and dermatologists, and this is very important. This is key. So we are the only one, um, in the market that identifies, you know, all the types of pimples. our AI was trained on ingredients. It's ingredients intelligence. So basically what it does is it, it scans all the ingredients of all the products of the catalog, and it picks the one with the best ingredients for your skin, as a dermatologist would do, right?
A good dermatologist, this is what they do. So basically, maybe you have, I don't know, uh, comedonic acne, and we give you, um, I don't know, a cream for [00:07:00] anti-age, you know, an an- with an anti-age marketing claim. We don't care about the marketing claims. We don't even see them, and that is very important for two reasons.
The first one is that, you know, sometimes marketing claims, they're not driven by science. They're driven by, you know, what the public wants to hear and how the public understands that, uh, cosmetic product, and that's, you know, part of the game. And the second reason why it's very helpful is that, uh, it m- this means that our AI requires zero effort for the beauty brand or the beauty seller because you don't need to, you know, um, manually adjust all your tags, product tags, uh, and to recombine it with the AI skin analysis results because our AI automatically reads your ingredients and picks the bes- the best products.
Deanna: I th- I think it is particularly interesting that your technology is trained to look for different types of this particular skin condition. I guess I'm wondering, [00:08:00] uh, can you say anything about the, uh, the sort of the training in terms of the images that were used or- I mean, were you looking at live people during the, sort of the initial AI training?
And, and then how does the, the platform continue to learn now?
Elena: Absolutely. So each of us at Dermaself is a, a skin specialist, meaning that each of us, every day, we do machine learning of our AI, right? So every day we keep doing that. So you really have to think about a human person looking at a picture and drawing, like, little boxes wherever they see a pimple, a scar, a spot, a wrinkle, and so on.
It's called the object detection- Mm-hmm ... in the AI language. And it's a- Sure, but a
Deanna: human is doing it first?
Elena: Yes. Okay. And we keep doing it. So, you know, again, y- you could use a machine. You could use, you know, AI and so on to do it, and do it in bulk, you know, and do, I don't know, 70,000 pictures per day.
That's [00:09:00] not our approach. Um, so just to give you an idea, um, when I do machine learning, usually it take me, like, at least half an hour to annotate one image- Mm ... and I do it manually every day- Mm ... at night, you know, in front of television. This is, this is my hobby. This is what I do and I love it. I'm, I'm a freak in this sense.
Uh, I really like it. So you w- um, so we identify pimple by pimple, and we, you know, we teach the machine which type of pimple that is, the severity, and so on. And we do the same with spots and wrinkles and enlarged pores.
Deanna: Okay, okay. So interesting to think about. I wanna think about a topic that I think comes up quite often in, in technology, especially new technology. And as, as familiar as we've all gotten with AI, I think it's certainly early days, uh, for, for artificial intelligence. I wanna think about ethics. Um, how are you... I, I hope you're doing it, so I'm just gonna say how are you-
taking steps to ensure that your platform is [00:10:00] inclusive? Can you, can you talk about that?
Elena: Absolutely. So we did a couple of software developments with, uh, Politecnico di Milano and Torino, which is, uh, an incredible university here in Italy, which I'm sure you know, you might have, uh, heard about. And one of these was to anonymize Pictures.
So it's very interesting. The way it works is we take the selfie whenever, uh, you know, a us- a user upload it in our platform. We take the selfie, and we kind of, uh, uh, cut virtually the face as if it were a mask. Uh, we cancel, of course, eyes, nose, and lips, and then we put this mask on a sort of, you know, uh, kind of statue, you know, like a fake face with standard volumes.
So basically, you keep the skin as it is with all its imperfections and wrinkles and pimples and so on, but you put it on a different face, so you are unrecognizable. Even the [00:11:00] most advanced AI today we have can't recognize you, and that's non-reversible. So we really care about privacy, and we look at privacy as our main competitive advantage rather than, you know, something that you have to do and sacrifice, uh, that you have to think about.
Deanna: No, that's so fascinating. I've... I do love that idea of anonymization of data, but also what you're describing, it's, it's so interesting because in some ways it's like further anonymizing it actually for the technology. Not only can they not fully identify the individual, but it's probably, I don't know if it's polite to say this, but in some ways erasing ethnic markers, right?
With the facial volume change that you have described, um- Yeah, it's so, so interesting. Oh, my brain is excited. Um, but I, I also wanna talk about, um, another sort of larger plan that I, I believe your [00:12:00] team has, which is to become a global beauty identity provider, uh, developing what we might call, um, a digital skin passport for consumers.
Um, what, what can you tell us about that plan?
Elena: Absolutely. So I think as an entrepreneur, especially in a startup, you always have to think about how the market will look like in two years, three years, right? So today, I look at the market, and I'm very positive, I'm very happy, you know, because of all the reasons I told you.
I think we have incredible competitive advantages. But I also know that in couple of years, you know, with all these, uh, venture capital funds and so on, all this money going to startups and beauty tech startups, and the market that I see that keeps growing, I'm sure that, you know, other amazing companies will come into the p- the game, into the picture, and I'm sure that they will, you know, look closely at Dermaself and start using, you know, similar technologies and so on.
So I also think that in couple of years, there will be a [00:13:00] lot of players like us, and that's when we will need to find a new competitive advantage. So I think about the consumer, and I think about a world in which AI is everywhere, in all the shops, all the e-commerces, and so on. Imagine being a beauty enthusiast like us.
Every time you buy skincare, you have to take a new skin test, and that's couple of minutes, right? That, um, kind of puts a distance between you and the checkout, the purchase. And then, you know, now we are all excited to use AI. But in couple of years, I'm sure, you know- ... we will be quite tired of it. So you will really need something quick, you know, um, a login, uh, to just buy your product, uh, or discover new products without having to do a two minutes, three minutes skin test.
So what we want to become is kind of like the skincare PayPal, you know? Whenever you go into a website, you visit a website, or you go into a store and you see the PayPal [00:14:00] QR code, you don't have to repeat your payment details because PayPal knows them. So you just have to log in, and, you know, here you go.
You can just pay with, uh, with your PayPal account. Same thing here with Dermaself. We are also, um, in the B2C software market, let's say. We don't monetize from it, but we have an app, a consumer app that you can use to take a skin test, s- track your changes, and so on. Track your skincare in general. So you will be able to use your Dermaself account that you have on your free app to log in whenever you see the Dermaself widget or QR code offline.
So, you know, Dermaself knows your skin. You don't have to repeat the skin test, and it will pick the best products for you in that store or e-commerce. Mm-hmm,
Deanna: mm-hmm. Yes. It's remarkable. Um, i- in my imagination, this project, uh, will be a, a data goldmine commercially, but I think also it will very much have real [00:15:00] implications in terms of w- what you're pointing to here, consumer freedom.
Can you talk more about what this will mean for consumers and, and maybe for other stakeholders as well?
Elena: I think actually the main game-changer thing we're introducing here comes from ingredients intelligence. Mm-hmm. So, you know, my, you know, my co-founder is a dermatologist, and when she looks at all the ingredients in all the cosmetic products, she is always astonished by how sometimes, you know, companies put into that plastic tube stuff that is not really skin-friendly, and she gets angry, and she's like, "Why are brands putting these ingredients in these...
on the faces of people?" You know? She can't understand it. And the reason is very simple. They just don't know how harmful they are because, you know, the law is quite slow compared to research- Mm-hmm ... and so on.
Deanna: It's so interesting. I, I'm sorry, I I'm compelled to jump in, but y- you're talking [00:16:00] about selecting products for the consumer not only based on the ingredients that will have positive benefits, but on selecting against products that might have detrimental benefits.
Yeah, that's the key here. Well, detrimental benefits, that's not accurate, but You know what I mean. Yeah, I,
Elena: I got it. I got it. Absolutely. That's the key. So basically, imagine a world in which, you know, consumers use Dermaself to find the right product. You know, who produce cosmetics will start to think, "Oh, wait, but when I develop a product, if I know which ingredients, uh, you know, are best for Dermaself, for the AI, then I'm sure that that product will be recommended more to the public.
So wait a second. Let me not put silicones in this cream. Let me avoid, uh, glycols, for instance, you know? Let me use something that is more green, more healthy for the skin because this way Dermaself will recommend it." So we're looking at a real shift here, right? Because you... they will need to talk [00:17:00] with a dermatologist.
They will need to get a dermatologist into the room whenever they develop, and I think that will be a great shield for consumers. You don't need to know how to read an ingredient list to avoid the harmful ingredient. You know, it should be by design. Mm-hmm. Whatever you buy, it should be best for you and your skin.
And we, all our skins, they're all different.
Deanna: Before we finish up, I wanna ask about the potential of this DermaSkin AI technology for clinical testing. You're looking at all of these images. You're doing skin analysis, which is certainly a, a key piece of, of clinical testing. Are you or will you be partnering with ingredient makers or, or product developers and, and looking at clinical testing?
Elena: Absolutely. So, uh, actually, our first B2B client was the San Francisco Research Institute.
Deanna: And r- just remind me what they do.
Elena: So they do clinical trials. Okay. Perfect. Basically. Yeah. Thank you. [00:18:00] Just to answer your question. And basically, they wanted our AI, um, to replace some human labor in their clinical trials process. So imagine you are testing out, uh, a, you know, a face cream, and when you do the clinical trial, you need people who review the pictures of volunteers at T0, T1, T2, and so on.
So, uh, after a month, after two month, you need to see the results, right? You need to measure them. Uh, you know, I told you, our AI counts the pimple and understand the difference between them and so on. So it's a great tool to understand if a product actually works. And that really helps, uh, whatever clinical lab to do that.
Deanna: Mm-hmm. Mm-hmm. That's so interesting. Excellent. You know, Elena, what you've shared with us over the last fifteen or twenty minutes is, is such a compelling, uh, and specific example of the promise of AI and digital data. I thank you for being my guest on CosmoFactory today.
Elena: I [00:19:00] thank you for having me. Thank you.
Your questions were brilliant, and I'm always happy when someone actually really understands what we do and doesn't, doesn't scratch the surface, you know, and really digs deep to understand why we're doing it.