A New Global Sizing System | Wardenburg Tailors | PI Apparel

So my name is Tessa Wardenburg. I'm founder of Wardenburg Tailors. I started my company 3 years ago with a mission to make perfect fitting clothes. We focus on people that find a standard very important. We focus on the leaders and the thought leaders of tomorrow, the designers of NASA, the first female director of a national bank, athletes, sports clubs, private plane companies, people that appreciate a perfect fit and demand this. So please raise your hand, whoever experienced this in store. Cool. So the first time I acknowledged the first time when I saw that this fashion sizing system has some few mistakes in it and has some room for improvement was when I started my first job in a little denim store in Antwerp, just across the Fashion Museum, when I worked inside. And I found out if I give 10 different pants to a customer, sometimes even none of them fitted. And I realized that this system is not user-friendly. This system that says everything to— that completely says everything 28, but in reality has maybe a 5% success rate. So eager to sell more, I started to really look at my product because we're calling it All 28, but every product, as we now see, is different. And I learned— I'm beginning to look at my consumer, and every consumer has different body dimensions. So I started my success by really focusing What is my product? Looking at my consumer and trying to find the match. Because as we all know, in store we only have 3 chances, 3 different paths before that customer gives that product back and goes to the competitors. So if you look at it from an outside, the success rate is only 5%. No other industry would accept such a low success rate. None. But we're still operating in the same way. So looking and diving into this, I made my success from this little denim store that had a turnover of 200K to almost 2 million. I could not explain what I really did because it was my human intelligence looking at the morphology of that person and mapping exactly the right product that matched that morphology. And then I could sell. It was my success of turnover that made me into a retail manager and expansion manager. I started to do the buying, and I wanted to understand who that customer was so I can buy exactly based on that morphology. I could not explain it why I was such a good salesperson. My fascination for this sizing system and this perfect fit grew. That's why I started my company. When I learned it, how to tailor, I fell in love with making the perfect fit. And there I grew the system. What are we actually doing with our sizes? How come do we accept such a big default business model that only has a success rate of 5%? So let's look at some numbers. If we're going into that, we're actually basing our whole industry on 3 averages. Here a small, medium, and a large. But in fact, every person is different dimension. So that means that these averages, nobody fits into them, because by definition, by mathematical definition, an average is nobody. Everybody is unique. So what are we doing? We're making our clothes on nobody. And then we complain we don't have profit margins and throw away almost 70% of what we make. What actually— and we know the statistics— we are one of the biggest polluters in the world. So what are we doing? Let's dive in some statistics. Today, An average is based on the highest density of people, excluding almost 80% of people. And funny enough, this is exactly the same number as my conversion rate in store, more than 80%. So what are we doing? So I dived into history. How come are we working on this default trial and error system? So around 100 years ago, we re-engineered the tailoring process. Where, for example, Dior for Air France started to make uniforms stating exactly the person that could fit into that uniform. The height, the weight, the age, the sex, the hips, the waist, the bust ratio. And we started to communicate that perfect fit. We are today still operating on that same system. We just replaced it by an average. But I don't know what's better. Every country, every designer, every fit technician made their own model. And we communicating our model to our consumer. This is not a user-friendly system. This is a system based on averages, based on people. So what if we could start unpacking this? What if we see that we have a huge opportunity, a huge untapped market of almost 80%? What if we can start unpacking this huge unpotential? What if we could look at all of our consumers? And not on a land average. Because let's face it, the morphology of a Gucci client is very different than a Primark client. So the whole concept of making everything on one is actually excluding everybody. So what if we can start designing on our real customer? What if we can include everybody? What if we can build our sizes on our actual body data. And maybe we can start making different models so that at least our whole collection is based on somebody, and preferably your client. So as an outsider, I recognize two big worlds. I recognize the production world— oh, sorry— where we speak about a garment And all of a sudden it becomes a size. And in a retail world we speak about a morphology, about a customer, and we speak about we sold so many mediums. But what are we saying? We are driving our car looking in the back mirror. We sold 50 mediums. That doesn't say nothing. And on the other side, we're speaking about the dimension of a garment. Both have huge amount of technology on it, but doesn't make a point. So I needed to prove my theory. So I changed 2 years ago my own fashion tailoring brand and convinced the Belgium national soccer team and federation, and I said, "I want to make the first collection based on morphology." So that's what I did. I learned from my friends that had different fashion brands how to make a collection. So we first made a mood board, we defined our target group, we know it was— we build a collection based on the football team, on Belgium. We design it, we're making our tech packs, our tech sheets, and then we started a new process. We started scanning the whole federation in, man, woman, athletes, board members, and we start looking at their data. And together with my co-founder, we started clustering the process. How can we go? Out of this clustering process, there came the different body shapes that I was going for. And we identified the footballer, the footballer, and the board member. This was the shapes of my real customer. And I had said to them, "I will make a perfect fitting collection that is as good as my tailoring that I had." So I did this process. I looked at the average of Belgium, and I saw my real customer had of course only 18% of them could fit into. So if I started my process based on the average of Belgium that I could buy from Elvanon, I could only serve 18%. That was not an option. I needed to make a collection that everybody fitted perfectly in. So I looked at their real morphology, clustered them and put them next to my first initiative drawing, my first tech pack. And in this case, I looked at the trousers, and as you can imagine, especially in the small and the medium, my upper leg was way too small. So by looking at their data at the beginning and the medium end, I could make their upper leg 8 centimeters bigger and I could calculate how long it needed to be. I did this process over the last 2 years with every product that I had made: t-shirts, pants, dresses. And we started also even taking it a step further. Could I make a gender-neutral collection? Could I say men and women? Can we maybe, if we start from measurements, if we start from data, Could we then design for both genders? And the answer: yes, we can. Very interesting to see. We did it with the polo when we started from real body data, and we clustered both of them together. What was a really interesting case was when we asked them all, is this fitting good? We saw— starting to see trends. Where we first looked linear, we now saw, we saw a trend coming. And that's a very good sign if we're talking about AI. We looked at the hip-waist ratio, which is really important in sports because you go different morphologies. I call it positive deformation because they trained very hard for it, but it's reflecting a reality. Of women and have very different body shapes. The results were astonishing. We could calculate that we can go from an 18% coverage where people could fit in to 89% coverage. Do you know what we're doing here? Do you understand that if you can go from 20% conversion to 86% conversion, of your customer that's coming in, wants to buy from you, convinced by the marketing team that they want to buy from you, and then only to be disappointed and crying and walking away. Now you can service 60% more. What that means in conversion, in sell-through, by this process, We could then also produce 40% less because now we knew our customer. We knew our shape of our customer. So by knowing that shape of that customer, we can alter that fit based on that shape. But that was not the first thing. We could calculate how much we needed for our target group. So we can easily know how much smalls, how much mediums. You know what 40% means in logistics, in raw material that we use from the earth? And that's ending up again on the same earth. And then it was really cool because of course we could suggest by 95% sure which product was for which person. So what we did, if we're looking at our heat map, we made sure that everybody could fit in. And of course, I know what you guys thinking: 86%, that's not everybody. It's true, we need to eliminate the outliers. We're still making fashion, we're not bespoke tailors, and hopefully we can eventually go there, but it's not a reality on the moment. But we can identify maybe an outside group, can make a model especially for them, or make it longer And of course, we needed to standardize the new sizes. So how can we make a system? How can we make a new system that's not based on an average where nobody's fitting into? So I identified again these two worlds, these two worlds that today are not speaking. So we started our moonshot a year and a half ago, where we said, "If we cannot standardize people because we're excluding 80% of our customers, could we shift it and could we standardize the way we measure our garments? Could that be a solution?" Because I know 100% sure We have a whole collection. So by giving them all the same name, we're actually confusing people. So what if we, if we cannot standardize people, what if we can standardize the way we measure? And is it not also our responsibility to share with our consumer what we made? How are we intending this fit? How does this collection must look like? What is that material? Is that not also our responsibility to start sharing? We're looking to see that we can make this industry user-friendly. Again, 95%, it's not a success rate. How can we make that better? So we started on the production side. How could we recreate a system that could be user-friendly for its end consumer? Of course, we needed data from that end consumer to start our new collection. So we made a very simple statistic model calculating the morphology to then start heat mapping it inside our designs to then go back and say, "Now please only show what fits me." Because as a consumer, it is confusing. You have 50 million options, but you have just one body. So can we maybe show only to that customer what actually fits? And now we're beginning to have predictive power. If we start from real data, how much quantity do we need? How does it need to be graded? How needs to be sized? And as well, imagine just saying to your customer, I made a very cool design, this would fit. Perfect for you. As a tailor, that's my secret sauce. I sell perfect fitting clothes. So what I'm not at is a designer, because you as brands will determine which color, what are you presenting. This is marketing. What we can do for you is make sure that it's sized and fit perfectly to your consumer. So we're not done yet. We're about 70% in our coding process and we need your help. So at the moment we are researching with the first brands for a 12-month period to start heat mapping and to see how production data and collecting body data to see how your collection is doing on the moment. This is not yet directly an overtaking. This is just the first step of unpacking your consumer shapes inside your collections. Over the last year and a half, we gathered a team with enthusiastic— that know that needs something needs to change. So we gathered the A-team of the industry of technology, AI, to start a new sizing system. That is inclusive, diverse, gender-neutral, sustainable, and let's make nothing more than what's necessary. Thank you so, so much for being here, and hopefully we can do something very beautiful together to make very, very good fitting clothes. Thank you so much. Are there any questions? Sweet spot? Thank you for that question. Exactly. I think the most important thing is that we as an industry have to play our part. We have to start sharing what we meant with that product. So the first step is not sexy. It's not AI or it's not that. It's just structuring product data. It's just setting the scene for machine learning, for an AI to work. Because I learned over the years that none of AI can work if you don't have structured data. It sounds nerdy, but if we all do something different, none of the technology can work. So that's the first layer. Do you encrypt that? I mean, if you get— do you get patents or drawings from— customers, brands, how do you protect that? How do you receive it? What do you do with it? So we've made a system that is very easy to build measurement sheets. So we ask the right questions, where we fill in, we preset how we need to measure in all different languages based on all 6,000 ISO norms. We also found that we are 6 ISO norms that are not matching, But it's very interesting. So you see, there are two worlds. The first step is to make sure that the way we measure is unified. And I think that's the first step that we need to do as an industry to make sure that we— how we measure our clothes is unified. We encrypt it to make sure we only share the information that the customer needs. We can share about the fabric, we can share, we can use it as a digital twin in the industry. Yes, of course, but we make sure that the patterns will stay company-owned because it's a very important thing that we're not sharing our IP, right? Is a question. How do you get the size pocket for the What do you mean? Who said the question from the size bucket? Ah, cool. I just thought, right, like usually you have your graphs with like hips to waist ratios and you have the smaller bucket, right? How do you get them bigger? How do you get them to cover more waist? We go point by point. So measuring point by measuring point where we layer the morphology, the clusters of the morphology, So we first do something that's called key main clustering to make sure that we have that, and then we heat map cluster to make sure that all the outliers are away, and we make sure that we put them next to each other. So if you make a decision how to make that fit, we know the intended fit. So point by point, measuring point by measuring point. Which means you just change the— Does it change the grading of the material? Or do you add the sizes? Do you also add sizes? Very good question. Exactly. We automatically grade it to that morphology that we had clustered. Yes. So we make sure that point per point, so it's nonlinear grading, did automatically. Something that one of our co-team members from Nike invented around 20 years ago, but that in an automated system, because this is a highly complex math problem what technology can do way faster. Where did you get the data from? So we made a very simple sizing app where we can ask our consumer 5 different questions and we have— we calculate their morphology about a data set of around 10,000 people and of course more, but it's to train them because I think every customer base is unique. So we need to start collecting data. So in the moment it's a very simple statistic model, but I believe as an industry you don't want your data stuck in a sizing app because 100% sure In 2 months there will be a new system to scan a body or to calculate or any AI. So what we need to do as an industry is share towards a new technology how to measure. And I don't think that we need to stock that in one sizing system because then that IP is shattered. And I believe looking at what we now do, we do single-dimensional tech. I call it old tech. It means your technology is good and how can you integrate with the rest of the industry. We have huge stakeholders, a huge value chain. So by going end-to-end is actually an old tech, learned especially from Asia where they do multi-dimensional technology. So it means you can integrate with different sources of information and start calculating on that. Collaboration is probably the only way how to make this happen. Is the idea to add more sizes? That's a very good question, and the answer: we don't know yet. We have to research your customer base. Do we see a trend in morphology? Maybe we can do less sizes. Do we see 7 different morphologies in a high percentage. Maybe we need more sizes, but I believe knowing exactly who we target, we could more sizes but still produce way less because we need exactly how to shape that. And that's the way how we can get a one-on-one relationship with our customer again. So do we need more sizes? Depends on your consumer. The population that we had was 100, around 100 people that we did with the Federation, and it was around 65% were male and quite a lot of also females. So that was the, the test group where I did it. Is the statistic on how much of the population? Anybody wants to clarify this question? Yes. So you were saying you went from 19% to 95% or something. How do you decide that the garment built to this spec will fit this person or not? That's mathematical. Of course, we have to find what we're researching with the multi-dimensional as well, the fit preference of that person, where we saw a trend in fit preference. So we're following that right now, but we need to of course look at what our customer, our personal fit preference, but it's something we can learn over time. What we can do is the mathematical, like a tailor, make sure it fits. The preference fit will something that we will learn later and later to reintegrate and to make the machine learning better and better. So it's something that we need to start learning from our consumer. I think it's also quite important, just also speaking to the big brands yesterday, of how you have Europe, you've got Europe, you've got Asia, and you've got Africa, and the body types, and even South America, the body types are completely different, but how do we work smarter going for the 80% and not just the middle 20%? Because it sounds like Africa is also just untapped still. Yeah, that will be speaking— thank you for that question. I was also speaking yesterday, even speaking to Alvin on a few years ago, there is no even standard for Africa there because the reason was they are too diverse. Very interesting, as we know, in 10 years we have more Africans on this planet than any other continent. So there is a huge untapped market going into this. Today they just use the Asian standard or sometimes the American standard. So it means a whole continent is underserved. So for anybody that's up for a big underserved continent, Please come talk to us. I would love to do some research and dive into the continent of Africa, as they are now everywhere. It's a global world, so there are body shapes everywhere. So instead of trying to standardize that, we need to come up with a better system and make sure that we can target, we can make our clothes fit to our consumers. And there we see trends, and that's the way we can start building around our consumer. I wanted to thank you so much for being here again and hopefully talk to you soon. And please come talk to me if you feel up for this challenge and see how we can get into this huge untapped market of body shapes, shapes created a new fit world. Thank you.