Perfitt: From Minimizing Size Uncertainty to Achieving “3-Zeros” Vision in Fashion Industry.

Online shoe shopping often turns into a battle of millimeters. A size too small feels constraining, while a size too large leads to discomfort. For shoppers, this uncertainty can result in blistered heels, cramped toes, or the hassle of returns and exchanges. For retailers, size-related returns are a logistical and financial burden, increasing inventory management complexity and lowering customer satisfaction.
Enter Perfitt, an AI-powered fit-tech startup offering innovative solutions to these age-old problems. By leveraging artificial intelligence and computer vision, it aims to accurately recommend shoe sizes, transforming the online shopping experience for both consumers and brands.

(Sunyong Lee, CEO of Perfitt / source = Perfitt)

What Is Perfitt?
Perfitt offers an AI solution called PerfittSize, which utilizes AI-driven computer vision technology to measure foot dimensions via smartphones. PerfittSize integrates seamlessly with brand websites, providing personalized size recommendations based on customer foot measurements and shoe data.
Brands adopting PerfittSize can reduce costs associated with returns, improve inventory management, and enhance the user experience through a user-friendly UI/UX, ultimately boosting sales.

What Inspired Perfitt to Leverage AI in Fashion?
Perfitt's story begins with my personal experience. Growing up, I was passionate about basketball, but finding basketball shoes designed for girls in Korea was almost impossible. My sister, who was studying abroad, would bring me basketball shoes as gifts, but they often didn’t fit properly. I still remember the painful blisters and even losing toenails because of those ill-fitting shoes. That’s when I started wondering: Is there a way to determine the perfect shoe size without trying them on?
Later, I launched a women’s shoe curation platform called Sugazine, and it became clear just how much of an issue sizing was in online shoe shopping. A significant number of returns were due to size mismatches, and I realized this was a problem that needed solving. Together with my co-founder and a developer, we spent three months working day and night to develop an AI prototype for foot measurement. That prototype eventually became what is now known as PerfittSize, the core solution of Perfitt.

What are the main challenges in the early stages ?
Building a startup comes with its fair share of hurdles. For Perfitt, early AI development required the team to measure foot sizes manually at various locations. Launching a global solution presented another challenge: Perfitt initially lacked sales expertise and had to learn everything from lead generation to pitching on the fly. Mentorship sessions with other global-facing startups helped refine these processes.

Were There Challenges in Collecting and Processing Large Amounts of Data?
Perfitt was founded before the era of generative AI, meaning the team had to learn the fundamentals of AI from scratch. The biggest challenge lay in identifying and processing the right data for the AI engine. In a fast-paced startup environment, the team worked in parallel to collect, verify, and analyze as much foot and shoe data as possible.

(source = Perfitt)

How Does PerfittSize Work?
PerfittSize uses a single foot scan to measure length, width, and shape, instantly analyzing the data to recommend the perfect shoe size. The AI engine combines foot measurements with detailed shoe data—such as interior volume, material, cushioning, and toe shape—to provide precise recommendations.
When integrated into a client's website, the Perfitt button appears on product detail pages. Customers can click to begin the measurement process, receive size recommendations, and complete their purchase—all without leaving the page. The tool also works seamlessly across multiple shoe models.

What is Perfitt’s Competitive Edge ?
Globally, size recommendation solutions fall into two categories: those based on past purchase history and those using current body measurements. However, relying on purchase history is insufficient for shoes, as even models within the same brand can vary significantly in fit.
Perfitt combines foot data and internal shoe dimensions to deliver unmatched accuracy. With a database of over 400 shoe brands and 500,000 foot scans, it is uniquely positioned in the market. Beyond PerfittSize, the company has expanded its capabilities to include VR experiences and live chatbots, further enhancing its offerings.

The Vision: Achieving Fashion’s “3-Zeros”
Perfitt’s ultimate goal is to help the fashion industry achieve “3-Zeros”: 0% return rates, 0% inventory loss, and 0% overproduction. Size-related returns account for over 80% of all online returns, complicating supply chains and increasing carbon emissions. Misaligned sizing not only erodes profit margins but also contributes to global resource waste.
By focusing on delivering the perfect fit, it tackles these issues head-on. Future plans include launching CRM tools to match unsold shoe inventory with potential buyers and using regional and demographic foot data to guide shoe production, reducing excess inventory and waste.
Perfitt envisions a future where fashion becomes more sustainable and efficient, one perfectly fitted pair of shoes at a time.

By Moon-kyoo Lee(munch@itdonga.com)