【DROBE Inc.】Saving "Fashion Orphans" in the 8 Trillion Yen Apparel Market. How "Selectbox" Pulls Offline Retail Shoppers into the Online Space via AI & Professional Styling
From Management Consulting to Fashion Tech: Saving the 70% of Women Who Love Fashion but Lack Confidence
Nice to meet you all. My name is Mamoru Yamashiki, Representative Director and CEO of DROBE Inc. I started my career at DeNA, later joined Boston Consulting Group (BCG) as a management consultant, and founded DROBE after recognizing the massive opportunities and structural issues in the apparel market. Today, I would like to introduce "Selectbox," our personal shopping service curated by professional stylists.
First, let me outline the service.
Selectbox is a service where users answer about 70 questionnaire questions on their smartphones, and a cardboard box containing five to eight items hand-picked by a professional stylist is delivered to their homes. Along with the clothes, the box includes a customized "Styling Card" containing detailed styling advice written specifically for each customer.
Our styling service fee is 4,290 yen per delivery. After the initial order, we suggest a planned delivery schedule based on the customer's lifestyle, allowing them to receive boxes periodically.
The reason we established this business lies in the extremely distorted structure of the domestic fashion market.
The Japanese apparel market has shrunk by nearly half, from its peak of approximately 15 trillion yen in the 1990s to about 8 trillion yen today. Meanwhile, the annual volume of clothing supplied to the market has grown exponentially. The market is flooded with cheap, mass-produced clothing, leaving consumers in a dilemma: "I don't know what to choose."
According to a survey targeting women, about 80% to 90% answered that they "like fashion," but a staggering 90% of those answered that they "lack confidence in their styling," and 70% stated that they "do not know what to wear, becoming 'fashion orphans.'"
In short, there is a massive silent majority of people who love clothes but struggle to select what to buy.
Let me clarify our positioning.
While massive fashion e-commerce malls like ZOZOTOWN or Rakuten exist, our target users are actually not the type of consumers who buy clothes on these platforms.
Approximately three-quarters of the Japanese fashion-buying population still purchase clothes at brick-and-mortar stores. This is because 90% of online shoppers fear that buying clothes online will result in sizing failures or that items will not suit them.
Although we are an online service, we specifically target these "offline retail shoppers who fear online sizing failures." Today, we have over 200,000 registered active users.
Empowering 300 Stylists with Workplace DX: Running PDCA Like an IT Company via Dashboards and Data
The strength of our service lies in our professional stylists who select clothes for our users. Currently, we have over 300 stylists registered with DROBE. Their backgrounds are extremely diverse, ranging from active stylists who handle styling for TV and celebrities to top store managers and sales associates who have mastered face-to-face customer service in major department stores.
We have brought dramatic innovation to how these stylists work.
DROBE's stylists work 100% remotely and on flexible schedules. Furthermore, rather than relying solely on intuition, they utilize granular data to run daily PDCA cycles, much like software engineers or digital marketers.
Specifically, each stylist is provided with a dedicated dashboard. It visualizes not only their individual sales but also detailed customer satisfaction metrics compared to the overall average, such as "size fit rate" and "design suitability rate."
For example, if a stylist's size fit rate for bottoms is below average, the dashboard doesn't just show a low score; it drills down to identify why. It analyzes whether the stylist tends to suggest sizes that are too large, whether they struggle with bottoms despite matching tops well, or if a mismatch occurs when styling users of a different age or body type.
Based on this data, stylists reference the best practices of top-performing peers to improve their next styling selection. This data-driven operation is what produces the high reproducibility of our premium personal styling.
Reducing Styling Time from 4 Hours to 15 Minutes: Our Proprietary AI and the Harmony of Humans and Tech
Since our founding, we have pursued the fusion of fashion and technology.
Initially, when stylists manually reviewed customer profiles and selected five items from tens of thousands of products, it took an average of four hours (240 minutes) per customer. This labor-intensive process was impossible to scale.
To solve this, we developed and introduced our proprietary "AI Styling Engine." As a result, we successfully reduced the time required for a stylist to finalize the items to be sent to just 15 minutes.
We do not sell in-house brand items; all our products are sourced from our partner brands. We currently partner with over 200 major Japanese and international brands, such as United Arrows and Onward, with more than 400,000 unique items registered in our system.
However, handling product data from 200 different brands presents a major challenge: data inconsistency. For example, even for the color "red," different brands register it as "red 05" or use their own color codes.
We developed a "data normalization system" that utilizes AI to analyze product images and text to unify data into DROBE's proprietary specifications. The AI automatically tags product characteristics (collar shape, sleeve length, fabric texture, silhouette, etc.) with high precision. This automated processing saves over 1,500 hours of manual data input annually.
Our actual styling workflow is a hybrid model that fuses human and AI strengths.
1. First, the AI automatically screens hundreds of candidate items from our database based on the customer's profile, past purchase data, and size restrictions.
2. Next, before shipping, we present these candidates to the customer online, asking for their feedback on which items they would like to receive.
3. Finally, a professional stylist reviews the customer's reactions (e.g., "I already own this," "I'd like to try this") and finalizes the five to eight items while arranging the actual coordination.
This combination of AI-driven efficient screening and human-driven emotional empathy is what realizes our outstanding customer experience and high purchase rate.
Returns-Specialized E-commerce with Onward's "Steppy" and Scaling via Influencer Collaborations
Let me explain our growth strategy and corporate alliances.
This July, we announced a strategic partnership with Onward, a major apparel company, to back their new shoe brand "Steppy" with a specialized returns e-commerce infrastructure.
Generally, offering free returns on apparel e-commerce leads to soaring logistics and inspection costs, ruining profitability. However, by implementing the automated returns processing operations and AI size recommendation technology we developed at DROBE, we built a profitable e-commerce infrastructure that handles returns seamlessly.
We are evolving from a single consumer service, Selectbox, into a B2B2C platform that solves the biggest pain points—online sizing and returns—for major apparel companies.
While Japan's apparel market is worth 8 trillion yen and the e-commerce segment is worth over 1 trillion yen, our current revenue is in the single-digit billions of yen. A massive white space lies ahead of us.
Particularly, to acquire "offline retail shoppers" who do not buy clothes online, we will strengthen alliances with apparel companies that have strong physical store networks nationwide. By merging physical stores with our online personal styling, we will provide a next-generation omnichannel experience.
Furthermore, as a new user acquisition strategy, we officially launched a select e-commerce service called "Android" in collaboration with influencers this year. By combining the social influence of creators with our AI data platform and styling backend, we will rapidly expand our reach to younger generations and diverse style segments.
We will continue to build a next-generation infrastructure that brings the right clothes to every "fashion orphan." We look forward to hearing from brands interested in alliances and investors who share our vision. Thank you very much.
Q&A and Feedback
Commentator (Mr. Fukutani): Thank you, Mr. Yamashiki. Having a professional stylist pick clothes and try them on at home, only buying what you like, is incredibly convenient. I struggle to find time to shop at stores, so I would love to use this service.
Currently, your registered users exceed 200,000. Do all of these users purchase items periodically? Please tell us about the actual size of active users and how they use the service.
Mr. Yamashiki: Thank you for the question.
The 200,000 registered users represent our cumulative user base. Within that, for our core subscription-style "Selectbox" service, we actively ship boxes to approximately 5,000 to 20,000 customers every month, depending on seasonal factors and promotions.
In addition to the Selectbox deliveries, we also offer an e-commerce function where users can purchase items individually online. Combining these, a large number of active users purchase clothing through DROBE on a regular basis.
Mr. Fukutani: I see. Tens of thousands of boxes are shipped and styled every month.
You launched the service in 2020. What was the marketing success or winning pattern that enabled such rapid growth in three to four years?
Mr. Yamashiki: The major growth phase occurred during the COVID-19 pandemic in 2021 and 2022.
Due to state of emergency declarations, major department stores and apparel shops closed, forcing retail shoppers to buy clothes online. During this time, our value proposition—trying clothes on at home with zero risk of sizing failure—resonated strongly with women who disliked standard online shopping. We acquired users rapidly through social media ads. The shift in the external environment served as a major tailwind for our initial growth.
However, compared to the 8 trillion yen apparel market and 1-trillion-yen e-commerce market, our current revenue is still very small. We believe digital marketing alone will eventually hit a ceiling, which is why we are actively pursuing offline alliances and B2B partnerships with major apparel companies.
Mr. Fukutani: So, to pull offline shoppers online, you are partnering with major apparel brands that own physical stores.
Another point: due to the nature of personal styling, there must be cases where items do not suit the customer's taste or size, resulting in a 100% return rate with zero purchases. How do you control this mismatch and return rate?
Mr. Yamashiki: Indeed, the risk of a 100% return rate is always a challenge. In fact, most similar personal styling services globally failed because their return rates were too high to cover logistics and operational costs.
However, we keep our 100% return rate (where zero items are purchased) below 20%, meaning over 80% of our customers purchase at least one item from the box.
This high purchase rate is maintained because the AI analyzes historical behavioral data and customer feedback from the pre-shipping stage to eliminate items that do not fit or match their tastes. Furthermore, stylists review their dashboards to learn from past mismatches. Continuing to improve this styling accuracy and reducing the 20% mismatch rate is a continuous challenge for our tech team.
Mr. Fukutani: While similar services struggled with high return rates, you have achieved a purchase rate of over 80%. I can see the strength of your AI and your stylists' data utilization.
I have high expectations for your future expansion, including influencer collaborations, and how you will hack this 8 trillion yen market. Thank you very much.
Mr. Yamashiki: Thank you very much. We will continue to drive the business forward.