Why I rejected "Scalability" to build a logistics nightmare (that actually works)
Hi, I’m Rosie.
When I started my fashion tech startup, Rosie Hong, every vendor told me the same thing: “Don't do custom. It doesn't scale. Just do Ready-to-Wear (S/M/L), manufacture in bulk, and run ads.”
They were right. The "SaaS model" of fashion (write code once, sell a million times) is efficient. But I realized that in fashion, "efficient" is often a synonym for "ill-fitting."
I decided to do the opposite. I built a business model that combines high-tech with high-touch. Here is the breakdown:
1. The Tech Stack (The Logic): We use 3D body scanning to capture the client's architecture. I don't write code to automate the whole process; I write algorithms that translate a 3D point cloud into a unique 2D pattern, correcting for posture and shoulder slope.
2. The Human API (The Soul): This is the part that scares investors but delights customers. Once the pattern is printed in Vietnam, the "algorithm" stops. We work with Master Tailors in Saigon - veterans with 30+ years of experience. They handle the "feeling" (drape, ease, stiffness) that math can't calculate.
The "Made in Vietnam" Arbitrage: For decades, my country has been viewed as a source of cheap labor for fast fashion. I'm flipping that script. I'm positioning Vietnam as a source of Technical Craftsmanship. We aren't competing on price ($5 t-shirts). We are competing on Dignity.
The Result: A garment that fits better than off-the-rack and feels more personal than a pure algorithm could ever produce. It’s harder to scale, yes. But the "Moat" is huge because few are willing to do the hard operational work.
Ask: Has anyone here successfully scaled a "Phygital" (Physical + Digital) service? Would love to hear how you handle the QC when humans are in the loop.
The hardest thing about B2B is that you're often selling to someone who didn't budget for your category. They need the result you provide but never planned to pay for it.
The products that win here usually create a new budget line (by being categorically new) or steal from existing budget by making the ROI comparison obvious. Which of those are you trying to do?
This is a really interesting trade-off — trading scale for quality and human expertise is often dismissed in startup circles, but it can create a much stronger moat early on, especially in physical + digital (phygital) businesses.
I’ve noticed that when humans are part of the core value delivery (craftsmanship, curation, service), the real leverage isn’t in software scaling — it’s in turning early feedback into repeatable operational patterns before you even think about automation. That’s a different kind of defensibility than “pure SaaS scale.”
Curious — how are you thinking about quality control and consistency as you grow? Is the human layer something you eventually want to automate, or are you embracing it as part of what makes the product valuable?
Another great insight, @mrHarsh. You pinpointed exactly what keeps me up at night: The Operational Patterns.
To answer your question about the human layer: We are aggressively embracing it, not trying to automate it away.
In the luxury segment, the "Human Touch" is a feature, not a bug. My goal isn't to build a "Robot Tailor," but to build a "Bionic Tailor."
Here is how we split the QC and Consistency:
The Tech (The Guardian): Algorithms handle the objective truth (measurements, pattern drafting, shoulder slope detection). It prevents the "human error" of calculation.
The Human (The Artist): Our Master Tailors handle the subjective truth (drape, ease, hand-finishing).
We use tech to remove the drudgery (the math) so the artisans can focus purely on the art. That, I believe, is the only way to scale "Soul" without losing it.
That’s a beautifully articulated split — “scale the math, not the soul” is a rare level of clarity.
The Guardian/Artist framing explains why rejecting pure SaaS logic actually strengthens your moat instead of weakening it. Most teams try to automate taste before they’ve even defined it.
Appreciate you sharing this — it’s one of the cleaner examples I’ve seen of scaling with human leverage instead of against it.