Interview: Fashion commerce startup Wishlink leverages AI to recommend clothes to users

The fashion commerce field is not one with a high entry barrier. However, it was definitely reckless to do ‘mobile fashion commerce’ when everyone wants to buy clothes after they try them on.

When such companies introduced their service to VCs, feedback was negative, but they continued with their service. This is the story behind the success of Kakao Style, a company created by the Wishlink team.

Although they were only a small startup, they moved to China in 2015 and launched Korean fashion commerce services until their business was overshadowed by the THAAD dispute. Before that, they had about one million users in the Greater China area.  

Wishlink, a company that has never chosen the easy path, recently released a new service, called One Day Ten Minutes. This service combines artificial intelligence and fashion commerce by ascertaining users’ preferences and recommending 100 clothes per day. Will they be able to change the landscape of the market once again? Below is an interview conducted with Min-Wook Kim, CEO of Wishlink.

Platum: You have been in the fashion commerce industry since you were a manager at Naver Knowledge Shopping. Is there any particular reason for this career path?

Kim: No, there is no particular reason. When I was working at Naver, I learned about online shopping mall advertising, and I think I learned more about how to make money in the process. I think that is why I continue to explore this field.

Platum: In 2012, you built Kakao Style with Kakao and you were in charge of operations until June 2015. At that time, the service was going well, and people were talking about the possibility of creating an IPO. Can you tell me more about that?

Kim: 2012 was a little bit unique in that people’s way of using the Internet was gradually moving from PC to mobile. No one was paying attention to mobile shopping for fashion. I remember having a meeting with one of the investment associates at a VC and she said, “Who would buy clothes without even trying them on?” That was only five years ago. We did not have any competitors then, and we made new attempts before others started. Also, at the same time, we were able to easily draw mass traffic from a platform called Kakao. We were lucky in some ways.

Platum: In the middle of 2015, you suddenly transferred the business rights of Kakao Style to Kakao. What was the reason?

Kim: We had some concerns. Although we were in charge of both development and operations, our identity was not clear because the service was named after Kakao. At the same time, there was a desire to run a business overseas apart from the Kakao Style service. On the other hand, Kakao wanted to make the Kakao Style service more sophisticated and internalized, so the interests of these two parties aligned. This is why we transferred the business rights to Kakao in June 2015. The money that we acquired during that process became the foundation of our business in China.

Platum: What kind of experience did you have during your two and a half years of business in China?

Kim: We prepared for our entry into China for about one year after the end of 2013, and we officially launched a fashion commerce service called StyleDo in the end of 2014. Doing business in China involves too many factors that we cannot control internally, because customs and Internet service policies are frequently affected by the government. In many ways, it is a difficult market for Korean companies, especially startups. The market is too complicated and difficult to describe in a few words.

Platum: In June this year, you released a new service, called One Day Ten Minutes, which combines fashion with artificial intelligence. It seems like another challenge for you.

Kim: When we closed our service in China, my team members and I decided on the last pivot and discussed new ideas for about 2 months. We could not build SNS or utility apps that did not have any relation to our past experiences, but since we thought we had the competitive advantage in the online fashion industry, we decided to try business in this field one more time. We thought it was important to differentiate our service from ZigZag, which is doing well focusing on online shopping malls, and Kakao Style, which is focusing on the advertisement business. As a result, the concept of recommending products to users through artificial intelligence technology came out, and so the current version of One Day Ten Minutes was released.

Platum: What kind of service is One Day Ten Minutes?

Kim: It is a service that makes it easy to find a fashion item that suits you with just 10 minutes a day. When you select the style you want and your age, the app will show you recommendations for the next few days, where you can say ‘no’ or ‘yes’. Based on this data, users will be able to find the fashion item they want and it will even be delivered to them.

Platum: Was the AI technology used in the service built by the internal team?

Kim: Deep learning technology, which has emerged recently, is not an easy area to utilize, but it can be learned quickly if existing developers are interested and study it for a certain period. AI technology itself is meaningless. The value of the service depends on the knowledge, data, and know-how of the company and how well the technology is applied. We are also using AI as a means of maximizing the resources that we have accumulated, in the specific field of fashion.

Platum: UI was simple and clear. Like Tinder, users can express their preferences by just swiping. To what extent is this personal data tracked and analyzed?

Kim: It grasps the user’s preferences through various methods; for example, by excluding products from the recommendation section, through the ‘like’ button, and with a detailed viewer. It is possible to analyze data, such as which shopping mall the user prefers and which product he/she finally purchased. It can also recommend products that have a high preference rating at a desired time.

Platum: Does One Day Ten Minutes receive commission from shopping malls? What is the profit model?

Kim: We won’t be getting commission at any point, and the main revenue model is advertising. Our platform has the advantage that it can be completely targeted. Currently, there are 50,000 unique visitors and 5,000 daily visitors. The retention rate is high due to the high satisfaction of the customers.

Platum: Who do you think are competitors?

Kim: In the end, big companies like Naver and Kakao might be our competitors. However, we have more experience and knowledge in the field of mobile fashion commerce than anyone else, and we applied deep learning technology as the main focus of our actual service. We will experience more trial and error than others, but that will make us learn more as well. This is our competitive edge.

Platum: What is the biggest risk factor for Wishlink’s growth?

Kim: We weren’t actually that young from the start of our business but I am afraid that my team members and I will pursue stability rather than new challenges as we get older. The older you are, the more realistic you become about quality of life and financial stability.

Platum: What are the short term and long term goals of Wishlink?

Kim: In the short term, we would like to get great feedback from our users and be recognized. This will be the foundation for a stable revenue model and then we can plan for the next steps. In the long term, we want to become a company that continues to create essential services for brands and users in the mobile fashion industry. Technology such as artificial intelligence is only a tool for achieving our goal; it cannot be a goal itself.

(Top photo from

This article, entitled “Interview: Fashion commerce startup Wishlink leverages AI to recommend clothes to users”, was written in Korean by Platum, edited by AllTechAsia.

© 2022 All Tech Asia. All Rights Reserved.

AllTechAsia is a startup media platform dedicated to providing the hottest news, data service and analysis on the tech and startup scene of Asian markets in English. Contact us:

%d bloggers like this: