Hi, everyone. Thank you very much for coming to my talk. It is a Berlin time 2:30 p.m. I have my afternoon tea with me. For those of you in different time zones. Please feel free to grab your morning coffee or evening snacks.Let's have a fun chat about the exciting challenges we're tackling at Zalando in online fashion world. So first just want to add a bit personal touch to my talk by a very brief self introduction. I was born in China, Beijing is my hometown. It is a city full of diverse thoughts and values. That's what I love and that's what led me to my series of adventure across the world. When I grew up, I have lived long term in three different continents, going from academia to industry, from media, medical research to technology, from finance to retail, e-commerce. I worked hands on as a product manager, analyst, data scientist, and engineer. I like small squad team, medium specialized team, large, full functional teams. The journey so far has been so colorful, unbelievably challenging and exciting. So when Zendo reached out to me, I was really excited, not only because of the Lando's core value and strategy, which was very attractive to me, but also the potential of joining a technology company that focused on fashion. Although most of my past experience are tech focused.

I always had this background influence from my mom about fashion. She was born in the time that everybody wore gray in China and fashion was not very accessible for most of the people. She had a passion to beauty, she cannot give that up. So she learned how to design and create fashion pieces herself. Her constant influence kept me interested in people's needs of this self expression through fashion. And when Zendo reached out, it gave me the opportunity to combine that hidden interest in me given by my mother with my next stage of career. So who am I to say? No, I joined the Zendo. That's all about me today. Now let's talk about the land of who is the lander. We are a leading online platform for fashion and lifestyle in Europe. As Europe's most fashionable tech company, we work hard to find digital solutions for every aspect of fashion journey, for our customers, for our partners and for every valuable player in the Zalando story. I'd like to share a few numbers to put the landlord's impact into perspective for you.

So here you go. We bring head to toe fashion to more than 48 million active customers. We have more than 7 billion visits per year. Customers can enjoy a variety of more than 5800 brands from world-famous names to local local labels. Our growth merchandise volume and GMV is about €14.3 billion. In 2021 we had more than 17,000 employees in Europe supporting 25 countries of customers and in fashion domain, we offer Europe's largest fashion assortment with around 1.4 million articles in beauty domain. We have created an assortment with more than 25,000 products and constantly expanding that inventory in 2021 we also partner up with Sephora to offer the most relevant assortment of prestigious beauty to our customers. It is a vast inventory and to help our customer navigate this vast inventory and enjoy the vast assortment. We also help our customer with fashion inspirations, size and fit devices, product recommendations and personalized discovery experiences. So landlord's vision is to be the starting point of fashion for our customers. But what is fashion? It is a seemingly simple question until you dig deeper. The answer is not that simple. That's why it's fun. There are many definitions of fashion. If you do a quick online search, the way different people act on their definition of fashions are even more diverse. So here is a Wikipedia reasonable definition of fashion for your reference and maybe serve as a starting point of our discussion today. Fashion is a self expression.

It is autonomy. It is for a particular period of time and place in the specific context. It can be about many things ranging from clothing to body posture. It is strongly influenced, even defined by trends. It needs to be available and popularized by fashion system to bring this complex complexity of fashion to life. I want to briefly touch on the Maslow's pyramid with you. Many of you may have heard the Maslow's hierarchy of needs in psychology. One of the great discovery, it divides human needs into five layers each field. On top of the other, the need range from the most basic ones such as food and warmth to the higher psychological needs of a feeling of accomplishment at the highest level. The pyramid describe a self fulfillment needs of self actualization. It is also very interesting that the basic and the psychological needs reduce when they're met. While the motivation of self fulfillment needs of achieving one's full potential increase when the needs are met, the more you get, the more you want it at the tip of the pyramid. So for some people, the simple use of clothing is to satisfy basic needs of worms to many others. It influence of psychological needs of esteem.

It consider if consider the definition of a fashion being a form of self expression, it could be closely related to the tip of the pyramid. The self fulfillment need and think about this very interesting characteristic of the self fulfillment need. The more you focus on it, the more you crave it. This is one of the major reason why fashion is exciting, fun and not easy. Self actualization is like that, exciting fun and not easy. It's not supposed to be easy. That is a fashion to different people. It plays different roles. There are many ways to describe people's fashion needs. And here on the screen, it's just one of them we use to help us better understand and serve our customers. Some of us have a higher confidence of style whereas some others struggle with it. Some of us see style as our identity. Whereas others style is a chore to deal with every day. There's a variety of people thinking about fashion differently to the pragmatist clothing may be serving the most basic needs want to. The improver fashion may symbolize their struggle to live to their full potential, just like fashion means different things to people. Different people may also have a different Zoo Lando in their heart. I would like to share briefly how the lender looks like in my mind. And to me, it is a platform that carries vast diverse inventory, helping different people fulfill their diverse need. And when you are ready, we are here to support you to explore fashion as your channel of self expression. It's accessible to you.

Potentially, we may contribute a small but very meaningful part to your pursuit of achieving your full potential, the tip of the pyramid. So let's pull back from this very abstract concept of Maslow's pyramid. So Lendo ground our support to our customers, different fashion needs in seven customer promises, we aspire to offer you flawless choice best in fashion inspirations. We tailor your experience to fit your interest. We provide convenience to you. We build trust with you. We enable you to make sustainable choices and we embrace what makes you you, those are our customer promises and they all sound great. But how can we fulfill our promises? And that's when technology comes into picture. When fashion Mets Tech, we make fashion not only exciting and fun for our customer but also slightly easier. Just like many other e-commerce, the lender face some common challenges. The lender has challenges such as a choice overload and long tail problems. The lender needs to be able to serve selections of product for the customer that blend the familiarity, novelty diversity and for new customers and new products. The lender has the same needs to face the cold star problem just like many other companies. These are the common challenges we're all facing as online ecommerce. At the same time, Zano has some very interesting fashion specific challenges. For example, as mentioned in the Wikipedia definition, fashion has a strong component of trends.

These trends are often fast changing with jargons that not easy to pick up for everyone. Fashion is also also very personal as well. Every one of us has a different body and different taste to make it even more complicated. Different brands may have different sizing systems and different fits, find something online that fits well is one of the most challenging thing in an online fashion. In addition to many people, fashion is not only about covering basic needs, the emotional needs of a higher level in Maslow's pyramid are very real in people's fashion life and many of us struggle with it. So can how can we overcome these challenges and make fashion more fun and less hard for our customers? So let know, is still on the journey to find the full answer rather than already achieved it. And just to be clear, what I want to do today is to share some learnings with you. So maybe you would like to venture into that world with us. The first exam example I want to share with you is called fashion driven browse. The concept is straightforward instead of letting customer read the different fashion jargons on our website, we use visual representations to communicate these concepts. Here is one example, the pictures show different types of colors. So you can choose the one you like by the look.

Even if you don't know the type of color is called Mandarian color, you know you like it or not by looking at it. This way is not only helpful for the customer to find what they want. We also heard the feedback that customers think it's even can be considered as an interesting learning tool for fashion languages and this type of visual representation can also be leveraged to introduce new trends and new styles as inspirations to our customers. And when tag come into pictures, when we use machine learning to verify the tagging of products to each of these visual represented categories to ensure a good customer experience. Another example are old friends, size and fit. We mentioned that size and fit is one of the biggest challenges for online fashion across the land of 25 markets return rate is about 50%. It is in line with the e-commerce industry, but it is nontrivial. Around one third of all returns are size related. Our sizing team actively address these issues through machine learning based sizing recommendations with our technology. We can also predict if an item will run small or big. Our measure, our measures allowed us to reduce 10% of all size related returns. And I recently had a personal experience.

I purchased a pair of shoes in a brand that I have never purchased before. I tried both the recommended size and my regular size. I decided to take the risk and go with the recommended size even though it was a tiny bit tight for me. But after a few wears, it turns out that these shoes stretch easily. If I went with my regular size, it will be too big to lose later on. I'm pleasantly surprised and impressed by the wisdom of crowd where our customer gave feedback about the fits and our machine learning solution, which provided me the suggestion. Although our size and fit team achieved a great success. We don't know the size and fit information about everything for some customer and products. We simply don't know yet. As you can see in the slides, I logged in as a new user in the similar item recommendation section. At least the two out of the top five product recommended for me doesn't have my side. It can be a disappointing and even frustrating experience for our customers for customer or items. We don't have enough size information. We could still filter for product recommendations based on size.

For example, we can use the size distribution of general population. We could also use machine learning to predict a less accurate size range and filter the recommendation results accordingly. From machine learning perspective, there are many things we can do. But if we take a step back to think more carefully about the customer experience, the desired behavior becomes debatable. We don't know enough about the customer size info, we may know even less about their intent. It is very common for people to shop for others rather than for themselves. Also, people can be a lot more flexible about size and fit than we thought. In addition, seeing an item that they like, although not in their size. So lender have the feature to enable our customer to get notified when their size is available. Customer may prefer to wait rather than never have seen this beloved product in front of them. So filtering items in the recommendation section may take away customers options such as these.

You can see there are many things we can do with technology, but ultimately, we develop them to create the best customer experience. Customer needs are diverse. The solution designed for them goes way beyond the pure technology possibilities and talk about customer needs. I'd like to circle back to the Maslow's pyramid. We mentioned earlier, customers fashion need often goes beyond the basic needs, addressing their psychological needs or even self actualization needs in the pyramid. For that, I would like to share an anecdote that may have spoken loud up for many of our customers. A few days after I landed in Orlando, one of my beloved colleague approached me a Tian. I'm so glad you joined us. Can you take a look at this? Why am I seeing some random weird stuff in the session of my recommendation? And there in the middle of the screen he shared with me on the third position. A gold chain is staring at me in between a t-shirt and a winter scarf. And then I took a second look at the screen. The session is named perfect pairings at that moment. It was pretty clear to me for this executive member in his fifties. The gold chain was not so perfect for his psychological needs, let along his self fulfillment, needs of self actualization. An experience like that is close to what we call the broken windows.

Our customer experience, just like many other things in the world has many imperfections, don't take me wrong. The difference is that for some imperfections, customer would just shrug and move on yet for some severe ones would, it would create a moment of frustration and it is so strong, strong enough to hurt our relationship. This feedback is getting closer or maybe it's already qualified as this break, broken window and it helped us immediately reve uh see some quite vulnerability in our solutions. The first one, what we realize is that we don't have an easy way to understand how representative this case even is. Is it just him or for many other customers? They feel the same way we collect signals such as clicks, purchase and likes from our customers. None of them give a strong signal of dislike. We could have millions of customers having this frustrating moments and we wouldn't never know that is a problem. Second, this feedback provider is a loyal customer who we can know a lot about through past history with us. Why we have recommended something that is clearly misaligned with his personal tastes. When we dig deeper, we realize the perfect pairing session is not personalized. It is based on the anchor product on the product page.

In this case, it is the sketch sneaker which is considered by many people as street style. That's why this gold chain come into picture. But for this customer, it is just a low key black sport shoes, comfortable to walk in. It's not about street style, it's not about the gold chain. So this leads to the last learning point I would like to share and mention about this case. We don't have to be perfect. We don't even have to personalize our recommendation. There's a certain charm of non personalized recommendations, but to build trust with our customer, we need to set the right level expectations. This machine learning solution we build can't live up to the promise of perfect pairing. This is where this small imperfect solution became a broken window.

So we decided to call it what it truly is air quote often wield together. That's what it is by setting the expectation, right? We build a trustful relationship with our customer. We provide value when we are helpful. If you are looking for straight style and coaching to go with it, we also get certain level of forgiveness when we meet the target. When people are looking for a very comfortable walking shoes to wrap it up last but not least it is a journey. Fashion is exciting. Fashion is fun. Fashion is not easy. We try our best to build a valuable customer experience which help with their fashion needs through technology enablement, make fashion more interesting, more exciting, more fun and a bit easier for our customers. This graph showing on the showing on the screen is an internal graph I share with my team. The major message highlighted here is that technology solutions, especially the machine learning based ones require a iterative process tightly combined with customer experience design.

Often takes many many iterations to deliver a good solution and a constant re evaluation of the customer experience with the technology is critical for the improvement. It is a long journey to be the starting point of fashion for our customers, but it is a very meaningful journey in our mind that conclude my talk. Thank you very much for attending. If you are interested to be part of our brief journey, please join us at our booth, join us at Zendo and let, let's take the adventure together in the online fashion world.

Sounds exciting. Thank you so much, Jen. This is a really fantastic presentation. I think we learned so much. I want to share the comment of Michelle. She she's sharing. I had never considered the link between Maslow and fashion. So some of the some of the insight that you share was really eye opening. Fantastic. And here is from Crystal for someone who struggles with fashion. I find this absolutely fascinating. So, thank you so much to end today. You did a fantastic job. And as you mentioned, the Lando is looking for amazing team members to join their team to work on the exciting projects to get together in fashion and tech. So, thank you so much and everyone who tuned in, let's give a works show a round of applause to Tian and make sure to drop by the Lando's booth. Say hi in the chat. Thank you so much, Tian once again and enjoy the rest of your day.

Thank you, Anna for the opportunity. Absolutely.

Our pleasure.