B to C EC conversion rate of the apparel industry is 11.5% in 2017 (Market research on e-commerce, Ministry of Economy, Trade and Industry (2018))。物販系分野全体のEC化率が5.7%であることに鑑みれば、アパレルはかなりEC化率が進んでいるといえるだろう。それに伴いEC独自の業務が発生したり、店舗では発生しない課題も生じている。

In the world of apparel there is a business called "sage" in industry terms. It is an initial letter of "shooting" "measuring" "manuscript preparation". Shooting filmed products as their name suggests. Measure the dimension of the product. Creating a manuscript is to describe the item. These tasks are fundamentally not occurring in store sales, and are EC-specific tasks.

また、消費者がECでアパレル商品を購入するときにネックとなるのがサイズの問題だ。S、M、Lや号数などの表記が一般的ではあるものの、ブランドごとにサイズが統一されているわけではなく、「こちらのブランドではMだがあちらのブランドではL」といったことは珍しくない。海外ブランドならセンチ表記かとインチ表記かといった問題もある。 A fashion tech company that tackles the challenge of such apparel EC is Meikipu. I talked to Mr. Tademoto, Representative Director and CEO.

ECサイトでも自分にピッタリのサイズがわかる

Two services deployed by Meikipu Corporation are "unisize" and "SASAGE.AI (cowpea ai)". まずはunisize。身長、体重、年齢、簡単なアンケートに答えると、買おうとしているアイテムのうち、どのサイズが自分にぴったりかわかるというサービス。主にアパレルECサイトの商品詳細ページに導入されており、料金体制はサイズ推奨数に応じた課金性だ。2016年2月からサービスを開始し、2019年1月時点で約70サイトで利用されている。

A site where unisize is actually used. Check the size using the unisize in the green part of the left figure. The result on the right is the result. Length feeling etc. are also displayed together.

ECサイトの商品詳細ページで、いざ商品を買おうと思ったときにサイズで悩んだ経験がある人は多いだろう。とくに初めて試すブランドだと店舗で試す機会がなかったり、海外ブランドだとサイズ表記が日本と異なっているケースもある。そんなときにunisizeがECサイトに導入されていれば、自分に合うサイズがわかる、というわけだ。

unisizeの使い方は次の通り。unisizeが導入されているECサイトの、商品詳細ページにあるunisizeのバナーをクリック。性別、お気に入りブランドサイズ、身長、体重、年齢、身体特徴(腕が長いなど)を入力すれば、自分のオススメサイズを紹介してくれる。洋服のシルエットも出るようになっており(上図右)、Lだと膝が隠れるけど、Sだと膝が出る、といったことも紹介してくれる。サイズのレコメンドエンジンは国内外に複数あるが、「丈感含めた着用感をシルエットを表示して教えてくれるサービスはunisizeだけ」(柄本氏)とのこと。

However, even with the same height and weight, there are some people who have long legs and arms, and it seems that accurate size can not be determined by just data such as height and weight. What is the correction there?

Unisize listens to height, weight, age and guesses the skeleton of that user. We got hundreds of samples of the body from various places, studied the structure of the human body, people such as height, weight and age are generally in such a body shape and what information about the length of the arm is the height It is to be understood. Further correction by listening to the physical characteristics there. Combining the information of clothes there, the knees come out and the feeling of length becomes like, and so on. Recommended for clothes. (Mr. Tademoto, the same below)

商品サイズのデータ管理に苦戦するアパレルブランド

unisizeは自分の体型にあったサイズをレコメンドしてくれるサービスだが、そのためにはユーザのデータはもとより、各ブランドの洋服サイズを把握する必要がある。繰り返しになるが、そもそもアパレルブランドのサイズ感は各社でまったく異なっている(だからunisizeが必要になるわけだが)。各アパレルのサイズは「腕の長さはxxセンチ、肩はxxセンチのトルソーを、私たちはMサイズと定義します。そこから何センチ上げた物をL、何センチ下げた物をSとします。」といった具合だ。これを各ブランドごとが独自に規定し、各社のサイズとしている。

In order for the apparel brand to introduce unisize, it is impossible to introduce unless the apparel side is grasping the size of each clothes. However, according to Mr. Tademoto, "There are many times that apparel is not grasping the size of clothes actually made in detail".

I was surprised to see the unisize business, but there were many phenomena that the apparel brand side did not have accurate size information of the product. It is supposed that there is no information as it is actually making the same item actually, but because production and sales are separated, at least the EC side people do not manage it as data It is said.

EC管理者が正確な商品情報をもっていない。そうすると、unisizeを導入してユーザに最適なサイズ情報を提供するということができない。そこでメイキップが考えたソリューションが「サイズを簡単に計測できるサービスを作る」ということだった。このアイディアが2つめのサービス「SASAGE.AI(ササゲエーアイ)」だ。  

The sacrifice cost is one third. Significant time reduction

サービス名にもなっている「ささげ」とは、「撮影」「採寸」「原稿作成」を意味する業界用語だ。撮影はその名のとおり商品の撮影。採寸は商品の寸法計測。原稿作成は商品説明の文章だ(ざっくりいうと、ECサイトに載っている写真や文章のこと)。ものにもよるが、通常のささげは1アイテム3000円程度で提供されることが多く、SASAGE.AIはこれを早く、安く、簡単にしていこうというコンセプトだ。  

(image: Mayukipu)

SASAGE.AI is a cloud service that uses AI to improve the efficiency of dedicated work. Hang the item you want to sacrifice with a hanger and shoot a picture from front and back. If you upload the data to SASAGE.AI, the size is instantly known.  

(image: Mayukipu)

When shooting, hanging on a hanger with a background struck by 10 cm unit dots. The size is measured by AI recognizing which part of the shoulder, bust, waist and hem are the parts.

撮影した画像は単に白抜きされるだけではない。3D技術を用いて、自動でモデルが着用しているイメージも生成できる。つまり一旦前後の写真を撮影するだけで、洋服の採寸もしてくれるし、実際に人が着用してるかのような写真も作ってくれるのだ。

Furthermore, SASAGE.AI also allows automatic document creation. For example, when you shoot a piece, you extract the feature points from the photo, this is a black one piece, the mold is mermaid, perfect for going out, and so on.

If you use SASAGE.AI, all the sacrifices will be done automatically. A general offering is about 3000 yen per item, but if you take SASAGE.AI, you can shoot, measure and manuscript everything you do with about 1000 yen, so the cost will be about 1/3. Also, normal souvenir takes about 1 week or so. With SASAGE.AI you can quickly create an image as if the model was wearing, and it is instantaneous to make a manuscript. So it also leads to saving time.

In the first place, there are few cases where souvenirs are entrusted to specialized companies, and in many cases fulfillment (a series of operations for selling goods in EC) is conducted. Basically, since apparel is going to deposit stock once in warehouses and logistics companies, it is most efficient for them to do dedication work.

Nonetheless, as a fulfillment company, work increases with the increase in EC sales, but a sense of speed is necessary and the amount can not be raised, so it seems difficult to handle offerings.  

For example, in warehouses, there seems to be a sense of crisis that it is not a case where shelves are moving automatically, and technologies like robots are being introduced more and more, it is not a case where we are doing inefficient task of sacrifice. So when I propose SASAGE.AI to a fulfillment company that is doing sacrificing work, I came to accept that it is cheap and quicker.

Labor intensiveness becomes a bottleneck of EC

柄本氏が仲間と一緒にメイキップ社を創業したのは2015年。当時のEC化率はまだ5%程度だったものの、これが将来15%、20%に伸びていくと言われていた。しかしそこで柄本氏らが疑問に感じたのは、「サイズがわからないのにECで洋服が買えるのか」ということだった。たしかにECは便利でこれから成長していくポテンシャルがある。しかしサイズがわからないということがボトルネックになれば、成長するものも成長しないのではないか、というのである。逆に言えば、サイズがわからないというボトルネックを解消することはビジネスチャンスである、ということだ。

In addition, Mr. Tademoto sometimes played rugby at school, "The size of the thigh is getting bigger in proportion to the waist". For that reason I had a chance to look for sizes matching pants, had an original experience with size.

In order to solve the problem that the bottle neck on the apparel side and the size of the consumer side are unknown, Mr. Tademoto founded Meikipu. We decided to start unisize business.  

アパレルのサイズ問題を解決しようと思っても、創業者たちはアパレルで働いた経験はありませんでした。なので社会科見学ではありませんが、色々なアパレル企業を見学させていただきました。そのさいにびっくりしたのは、このデジタルな時代にアナログな作業がなんと多いことか。すごく労働集約的だったんです。 しかもお話を伺うと、人が全然足りないという。かといって給料を上げる余裕もない。この状態からEC化率を上げようと言っても、絶対にムリじゃないかと思いました。だからこそテクノロジーを使った業務効率の改善が入る余地を、大きく感じたのです。

And then it was unisize that Meikipu first developed. However, as mentioned earlier, offering work is heavier on apparel in developing EC. Even in the sense that it solves obstacles to unisize deployment, I have also developed SASAGE.AI.  

Tackle industry challenges with technology and unique ideas

The function of SASAGE.AI has not been completed as of February 2019. As for measurement, although it can cope with various categories, one piece is mainly concerning model image generation. Other categories will be rolled out in order from 2019. As for manuscripts, it is difficult for AI to make perfect sentences yet, and it is necessary to make modifications by hand. However, I would like to raise its accuracy. If the number of products appearing in EC at SASAGE.AI increases, unisize will become a better circulation that more turns will be added.

Basically, unisize and SASAGE.AI are services that try to solve problems in the apparel industry and to streamline operations. Meikipu is a culture of trying to do what our other companies do not do with our own approach. I found an assignment, but no one is doing it, then it is a feeling that we only have to do. It is troublesome and troublesome around the size of the apparel, but it is a place where nobody has done. As for sacrifice as well, if some company wants to do similar things with a tremendous technical power, it will be brewed. Just to investigate variously, in fact, no one was doing, and the task had room for improvement, so we began working on it. I will continue to have such challenges in the future, if anyone does not do it, and I have something that can be solved with unique ideas, I would like to do it.

The EC conversion rate of apparel has been steadily going forward and will continue to expand steadily in the future. On the other hand, the labor force in Japan continues to decrease. I can not afford to allocate personnel to work that does not produce added value or inefficient work, and there is no way to begin with anything in the first place. Therefore, using the newly emerging technologies to think whether you can improve the efficiency of your work is also essential in terms of national growth.

On that point, Makekip is a company that solves apparel offerings and issues of size using technology, and will continue to work on technology issues using the technology. Although we are not making outstanding from consumers nevertheless, there is a possibility of becoming an indispensable service for the sound development of the apparel industry. If you have the opportunity to buy clothes this time at EC, please check if unisize is introduced.  

The contents of the interview are also distributed by podcast

We are delivering interviews at the podcast interview.