Watch the recording of the talk here
[00:00] Very fun being here today. It's a great pleasure to be introduced by Solvang. It's all top. My name is Andreas Kristiansen. I'm a commercial manager and co-founder of Villoid. Before I started in Villoid, I was the market manager in Miinto, and I was on a very good journey there, from 10 to 100 million — which I am very proud of. In the next 20 minutes, I'll tell you what happens when a bunch of people without professional experience get into the fashion business. When we got into the market in 2018, it looked like this: we had Zalando, Boozt, NA-KD, Nelly. They were the icons of the industry — extremely large companies that had built brands and companies for years. They had everything we didn't have: money, big teams and lots of customers. It didn't feel exactly like an open market. The competition was huge. Nobody thought we'd make it.
[01:22] And so we felt the Norwegian market was: we had big giants in the top right corner, extremely large, and the Norwegian shops were very small. There were many in the industry, but only a few large ones took an incredibly large part of the cake. Zalando and Boozt stood with a suction inside the Norwegian market — all the money, all the jobs, all the profits went out of the country, while in Norway we fought for the little ones. I think very few people knew how much Zalando and Boozt actually sell in Norway — it's about billions. When we look at our big online shops in Norway, we talk about Belakaker over a billion, Inspired Group over a billion and something. But they're small compared to Boozt and Zalando.
[02:35] And it's in the biggest category of the consumer — clothes and fashion. So when we looked at this, it gave us a powerful boost. We thought, can't we take up the fight against the big giants? We had experience from tech, from startups, from McKinsey, from online commerce — what could go wrong? We also had a lot of confidence, as you can see in the picture here. But there was only one problem: we had no professional experience. This is a typical Google search from my PC in 2018: ‘How to sell dresses? What is the difference between our and harvest collection?' It was pretty clear — we didn't know what we went into. We were supposed to be at Fashion Week. I had never been to Fashion Week before, came there in the skiny jeans and hoodie, and Jeanette decorated us a little at least.
[03:49] But we literally felt like we were out riding. The fact that we're coming from outside, without professional experience, gave us an opportunity — we could think differently from the industry. We knew nothing about the established, and we believe that we were our greatest advantage. If we are rewinding: from the start of 2018, we have gone to over 500 million on the market. There has been growth all the way, since day one — we have managed to create a growth machine against the big giants and take market shares day by day. But we haven't just created growth, we've also become profitable. From 2023 we have managed to make money too. We are very proud of that — the combination of growth and profitability is not many who can handle. So the question that someone probably has is, how have we done that?
[05:11] To understand that, you have to understand the market we went into. The quote here from Tony Robbins describes the feeling we had of the status of the fashion industry when we went in: many operated in the old way, rigging was too physical trading, very little focus on the web. Physical and online trade are two completely different things, most of them probably know in this room. The point was: we got into an old business, and there were especially some things that got out. When we were to do purchases, it was that hit us early that it was controlled by everything but data. It was a lot of gut.
[06:15] The brands had their own agenda, little contact with the final customer, and very much was guided by major trends from Fashion Week in Paris, Copenhagen and so on. I have a story from a purchase I was on in 2018-2019. We made a little budget — probably the size of a trial room. It was nothing. We brought our 30-40 thousand to this mark, far below their minimum limit. They said we have to buy at least a hundred thousand to make it worth it. But we brought the money we had, we didn't have any more — so we were completely dependent on hitting with every product we bought.
[07:06] We meet a seller, I had done some analysis in advance, googled and looked at search numbers, so I had pictures of what I wanted to buy. When I show them to the seller, he says, "That's cool with those numbers, but that's the last year. What you're going to have now are these shoes here» — shoes with roses on. It draws a picture of how disconnected the sellers were from what the final customer actually wanted. I don't think it's because the sellers wanted something bad — they just didn't know what actually worked against the end user, because they sold to physical stores with little data. Another thing we noticed was pricing — the campaigns in the market.
[08:06] It was marked by some rules: sales twice a year. Then you went from zero discount for a long, long time to dump everything at 70 percent. What you don't take into account is that a summer collection is long — you might get the goods in March, and sell them out in July. What you don't think about is that the goods you put on sale in July have been in the store for a long time. You have not sold anything, but you still have to have the summer sales in July, so you put everything at 70 percent even when the product is really still attractive in the market. It didn't make any sense to us. What we saw was that this led to overproduction — they ordered too much goods because they guessed the purchase instead of using data to analyse what the customer actually wanted.
[09:13] And then you dump it all at 70 percent in long periods. It's just bad business, and it didn't make any sense to us. When you shop online, you have data, you have tools — it is possible to solve this. We didn't understand why no one worked in another way, so we decided to attack this from a different angle. Back in 2018 we came from tech — that's what we could. Jarle, who is a member of the Villoid community, was very concerned that we had to use data and build an infrastructure that was completely raw. So we hired a doctorate in machine learning and AI, Tim. Together they sat down, work, ate a little strange, drank energy drink and built a state-of-the-art data sheet.
[10]:26] A data sheet is a large database where data is collected from all sources, standardised and ready for use across platforms and analyses. We were very early in building this repository. There was only one thing missing: data. So we had a state-of-the-art repository, but zero data points. Fortunately, the job we did was not lost — in 2026 we were over 3.2 billion data points. We saved every click, every purchase, every journey, and data from many different platforms, all in the repository. But I found a quote from Todd Park, a former CTO in the United States, saying that data alone is wasted — it's only useful if you can use it properly.
[11]:42] I think most people recognize themselves in that statement. You're sitting with a lot of data, you know, you have numbers in all the platforms, you have dritmy data as well, but what do you do about it? How do you make it add value to the company in the daily course? We started working on this, and over the years we've developed a smooth tech stack. We've run a best-of-bred tank, with many different external platforms that are flexible — we can replace at any time and optimize our tech-stick. From these tools, we're going to get data into our repository. We have also built self-developed pricing technology, to see what price a product should have at any time. We have optimized marketing.
[12] If a product is empty for the best two sizes — should you really spend money on Google to market that product? We've looked at demand precision, analysed profit per product, and in 2022 we decided internally that we should go from just rocket growth to also have profitability. This became absolutely crucial to us. We analysed the profit per product, cut unprofitable products, looked at the marketing budget and put into effect all these things. Data has been part of our DNA all the time, and we've used it to create real effects. And then the best thing that's happened: AI.
[13]:59] When AI came, we felt that we really got paid for the job we had done with our data structure, all the tools we had made, all the knowledge we had about how to use data. AI has really revolutionized the way we work, and how we work with technology development internally in Villoid. Now it looks a little bit different. We're still working in a pretty similar way, but we have connected tools — external tools like Shopify, Vertex, we're working with multiple AI tools that you can just connect to and use to perform actions based on data. We also have something we call co-developed AI tools — relatively young tools built up with AI, where we can tailor with the supplier to adapt to our needs. But the coolest thing is our own AI tools.
[15:11] We have gone from the point where there was a backlog in the tech department when you were going to make something new — which could take a month or two, and which meant that you had to give priority to it, maybe scraping some projects because it took too long — to be able to develop most of the things we want, in days rather than months. We've created planning and forecasting tools. For internal use only, we have created a sort of HR tool for self-development. We're making AI campaigns and product images. The point is there are countless possibilities, and we've made a lot of good stuff in recent years. To give a concrete example: we're in fashion, and a large part of fashion is content. This was a nightmare — we were going to have a photo shoot.
[16:26] Then you had to think about what products we were going to bring, people were sitting and lightening the numbers to figure out which products we were going to be taking part in the shooter. You had to order from the warehouse, book model, create the moodboards — and everything was distributed over two Asana boards, four spreadsheets, and a lot of communication on Slack and Gmail. It's history now. We've cut out all the tools, streamlined the process, using AI for better quality of what we do, and saving extremely much time. What's important about AI is that the possibilities are there, but don't forget that AI is not just a tech project. It's very typical in a company that you have an IT department that sits in a corner — no one talks with them.
[17:31] You go there either to complain or to make sure things go too slow. We have felt that there has been a kind of additional service in the company before, and that is something we actively have worked to avoid. In 2025 we all felt a breakthrough — there has been a democratisation of tech. Everyone can develop, everyone can use AI. But what's the key before you can really use AI for your success is the culture of the company. On the left hand side, you have a classic setup: some tech developers make long backlogs, it takes forever. I think people who work with tech and who are interested in it are throwing themselves over AI — they love it.
[18] But the rest of the company is a little bit like this: 'Ach, AI, I can't do that. No, AI is something for those who are doing it." But we believe very much in the power of engaging the entire company in developing and solving their own problems with AI. We set ourselves a goal with Jeanette and the leadership team that everyone in Villoid should be good at using AI in everyday life. Right now, vibe codes 9 out of 10 people in Villoid. We have Gunnar in the warehouse — he's sitting with a reception, and he's automated himself. It's wild: he codes and pushes out tools to streamline stocking. The Piid Social team has done something with Smartly, for those of you who know what it is — he has automated his entire workflow. The examples are countless, and this is where the gold lies: in getting the entire organization to use AI.
[19:48] For us, it looks like this, a little bit simplified: you need data to make the AI not just a fancy Google substitute. And to get exponential growth in the development of AI, you have to bring the whole company, and a culture on top that encourages employees to use it and facilitates them to develop and become nice. But we haven't just built a computer company — we've also become a fashion company. So what happens when a tech company goes into the fashion business? We've gone from riding, literally, to getting a little busy with fashion too. We're on tuxedo, we're on Fashion Week. We've got that skill in fashion too, and it's important to say that the fashion element is extremely important to us.
[20:57] From Google how to sell dresses, we are now aiming at the billion next year. It's not so bad for a bunch without professional experience. Thank you.