Cloth AI

A startup initiative to reimagine how we shop for clothes — 10.7% conversion rate, 80% eager to try app, 60% comfortable paying subscription.

Startup

Client

Cloth AI (myself)

Industry

AI, Fashion

Duration

2 months

Results

📈 10.7% conversion rate on fake door marketing site
📱 80% (12 of 15) eager to try app
💳 60% (9 of 15) comfortable paying subscription

See The Cloth Prototype

The Problem

Shopping has always been a struggle for me. Wasting hours at a mall sifting through clothes only to end up with pretzel bites is a defeating feeling (the pretzels rocked, though). Thankfully, I'm not alone in this.

Based off a NYP poll in 2022, 52% of americans struggle finding clothes that fit. That's millions of people. But in my experience, fit was only half the battle. I not only needed clothes that fit, but also that matched my skin tone, complemented my features, and worked with the other clothes I had on. To see if others had the same issues, I had to talk to real people.

The Discovery

Within my network, I spoke with five people to better understand their shopping experiences. What I found was consistent: shopping felt like a time-consuming gamble, requiring commitment most can't afford. Most didn't even want to shop and felt it was a necessary evil. It got me thinking. What if you didn't have to shop at all? What if someone or some "thing" could do it for you?

At the end of the interviews, I pitched this concept idea to each participant —A virtual stylist that did the shopping for you by understanding your body type and preferences. Everyone was interested. Some even said they were willing to pay. I was ecstatic, but I couldn't rely on 5 people telling me my idea was great. I needed unbiased data.

The Demand Test

To get unbiased data, I created a "fake door" landing page with one goal. Convince people this was a real product and see if they clicked "download app". This would tell me if there was demand for this idea. Ads were pushed to various social media platforms to get traffic into the site. After 3 weeks of testing with 2600 users, I garned a 10.73% conversion rate. Enough to tell me there was strong interest.

📈 2,600 visitors in 3 weeks
✅ 10.7% conversation rate — clicked "Download App"
✍️ 11 signups for waitlist

Designing & Recruiting

The demand was clear, so I didn't waste any time getting to work on running through various flows and prototype concepts. The idea of this experience was to first capture my user's preferences using a quiz format and then translate their data into an output of clothing recommendations. This process would only take a few minutes to capture their body type, skin tone, style, and fit preferences. No more having to guess what looks good and spend hours in a store as the virtual stylist will do all the work for you.

Meanwhile, I ran ads to recruit prototype testers and offered a reward for their time. It was challenging to find genuine testers and not just money grabbers, so I tightened up my screener to improve the quality of my participants.

The User Test & Results

15 users walked through the prototype and the overall reception was very positive. 80% said they wanted to try Cloth. 60% said they'd be willing to pay for it. Some even asked for a release date. All great signals, but I knew that true adoption would all be dependant on the accuracy of the body scanning technology and the recommendation algorithm.

💬 80% wanted to use Cloth.
💳 60% said they'd pay. (if the body scan & algorithm worked)

See The Cloth Prototype

One of many users giving their overall thoughts on Cloth. Permission granted to share publicly.

Should I Build This?

Testing proved this idea was sorely needed, and I was beyond excited to take this from concept to reality. To get grounded with the technical feasibility, I reached out to developers and AI research consultants to get their perspective. I learned that building the core functionality would require $300-500k+, a specialized AI team, and potentially 3-6+ months just to nail down the body scanning alone. This wasn't ideal for me as I was hoping to deliver most of the functionality within 6 months tops. It wasn't realistic.

Due to the technical reality check and how quickly AI was developing, I unfortunately decided to pause the project and shift my focus to something more predictable. The stakes seemed too high for me at the time. If I decide to revisit Cloth, I’d narrow the MVP to a single feature and validate from there — starting small, testing fast, and getting technical input sooner. This was a valuable lesson in market validation, scoping, and the realities of initiating a startup.

Cloth was received very positively, and testing showed that people wanted it now. I wanted to meet that demand, so I spoke with a few developers and AI experts. Given the scope and skill required, the project was going to take longer than I had hoped, and costs were projected well over $300–500K. I wasn’t ready to make that kind of commitment given the uncertainty around finding specialized AI talent and how quickly AI technology was evolving.

I ultimately decided to pause on this initiative and divert my attention to something more predictable. My job. If I were to revisit this concept, I'd lower the scope of work and zero-in on one core feature for MVP. Until we meet again, Cloth, if someone hasn't built you already.