See the original story in Japanese.
Start Today (TSE:3092), the Japanese company behind online fashion marketplace Zozotown, announced in January that it has set up a project team called Start Today Research (Research for short) as its R&D arm. Yuki Kanayama, CEO of Vasily which was recently acquired by Start Today, will lead the arm and conduct projects leveraging big data owned by the fashion giant.
The specific Big Data set as its analysis data include 30 million brand-official product data (product information, category, size, material), 10 million fashion coordination data, 23 million user information (user attribute, purchase history, held items), shop information and distribution-related data, in addition to human body dimension data measurements from Zozosuit as had recently been announced and is seen being accumulated in the future.
According to Kanayama, although he cannot disclose details about the current team, he aims to build an organization a few hundred-strong comprising doctorate holders in such fields as machine learning, encryption technology, material science or hydrodynamics, so as to analyze fashion from the scientific perspective in the future. In tandem with the announcement, we held a short interview with Start Today CEO Yusaku Maezawa and Kanayama, inquiring into their objectives in establishing Research [interviewer questions indicated in boldface].
Approaching fashion from scientific angle
I give due warning; the whole picture of the R&D arm is still unclear. As mentioned above, it is a fact that the firm has been storing an enormous amount of data related to fashion especially in Japan. The data content will enable more multidimensional analysis by adding body dimension data measured by Zozosuit that astonished the world. The two guys explained about the outline of the new arm.
Although Research was introduced with a bit ambiguous expression as “scientific approach of fashion utilizing big data” in the press release, would you explain about the outline of it again?
Kanayama: The purpose of Research is to utilize fashion data including human body dimension data. We will jointly research with external companies or educational organization about what can be grasped from fashion data or how we can utilize it into our service. The output of Start Today is thought to be divided into two types: the one is academic approach via papers or learned societies and the one is to make it into services or products led by companies. As acquiring an enormous amount of data, I would like to improve the state of the world together with outside collaborators.
Is it a R&D department of Start Today?
Maezawa: We had sometimes conducted planning or development as needed, but here we will cover a wide range of R&D activities in Research.
I ask you more about the organization. While Mercari had also announced the establishment of R&D department with detailed research themes or organization scale, how about is that in Research?
Kanayama: We plan to construct a several hundred-scale organization having Ph.D. members, in addition, we will start researching about fact or theory which has not been commercialized. Fashion has both “protection” and “emotion” aspects. The protection aspect is to wear clothes to protect himself against the cold or to cover up himself. The emotion aspect is to lift the spirit or to gain confidence or satisfaction. To meet the emotion aspect, measurement technology is needed, and the size suit must be improved further. We will explore themes about combination or coordination utilizing algorithm, machine learning, or deep learning. If health factor were required for fashion, healthcare field would become one of our research targets.
These data, Kanayama compared to “petroleum”, is just a string if it remains as it was. Refining it and synthesizing it, they turn it into a form available. The direction Research pursues may be to quantify fashion that had been determined based on the sense and to change fashion into a knowledge that everyone can share as common recognition.
Perfect-fitting by one-centimeter pitch
What are the specific views the two guys are going to cultivate, while it seems to a big project they imagine? I continued interviewing.
I will ask you about the research contents. For example, Amazon has a simple view like Echo Look, suggesting fashion which suits each user by taking photos. Do you have any specific ideas as outputs from Research?
Kanayama: Yes, we have some plans for services but cannot reveal them now. I want to pursue the point what is cool for him seen from himself just as seen from those around him.
Maezawa: The purpose of Research is to seek the answer to the problem as for what is cool for each user, by suggesting satisfactory fashion that looks cool seen him and also seen from surrounding people. We aim to digitize what is cool and what is cute. Off course, it is important for a user to be satisfied himself, but we want a proof that other people will be sure to admire the fashion enough to recommend them to him.
We will digitize the ambiguous concept of “cool”; for example, a coordination is cool because it is formed by combining this one and that one it is highly evaluated by 85 in 100 people qualitatively. Anyone may be worried about whether clothes will suit him. I think it is important to tell him that it really suits him or looks cool qualitatively to make him positive. To do that, first of all, we want everyone to experience wearing perfect-fitting clothes and to be impressed with them. Then, digitalized data outputted from Research will work as a support. That is our future vision.
As hearing Maezawa’s explanation, I had a dim impression that standardized information related to fashion like “fashion score” will be born from here. Of course, that will be incorporated into the group’s services, but standardized data can be used by external companies too.
Let’s continue the interview. If you could acquire any outputs from the data in Research, how the world will change? For example, with the advent of fast fashion represented by Uniqlo, we no longer have difficulties in having clothes just to live. How are they going to change the world?
We no longer have difficulties in having clothes thanks to the advent of fast fashion. On the other hand, ocular changes have occurred in the street; homogeneity in fashion can be seen. What impacts do you expect Research to have?
Maezawa: I think the silhouette of aman wearing T-shirt is an art; it completely differs according to person. In autos’ body lines, there is a ratio called golden section which is digitalized to a certain degree. Similarly, we will try to sublimate the silhouette in a size suitable for each person to an art. I believe there is “a point” of perfection.
Currently, we have been producing T-shirts with size intervals of several centimeters pitch and wish everyone to find their favorite silhouette. An impression of clothes changes greatly only with one-centimeter difference in size, but there had not been such line-ups allowing customers to choose thus far. Few people are particular about one-centimeter difference in clothes, right? That is why I wish them to find their own perfect T-shirts.
Although I cannot refer to Zozosuit because I have not experienced Zozosuit yet, how does it feel to wear a perfect T-shirt?
Kanayama: It makes you positive.
Should the day arrive when we really can obtain perfect-fitting clothes, we will surely get to enjoy fashion from all over the world via online more easily or more people will gain confidence about one’s own fashion by utilizing the abovementioned Fashion Scoring.
Will it bring another change just as fast fashion created? I would like to wait for their outputs with great interest.
Translated by Taijiro Takeda
Edited by “Tex” Pomeroy