Designing 0 to 1: What I Learned Building a Product from Nothing

Twice in my career, I've designed products from zero. No existing product, no legacy screens, no established patterns. Just a problem, a founder's vision, and a blank Figma file.

It's the most exciting and the most dangerous kind of design work. Here's the framework I've developed.

Week 1: Learn the domain, not the UI.

I once jumped into designing fintech screens without understanding how stock brokers actually work. I had to redo most of it. Now, my first week on any 0-to-1 project is pure learning studying the existing workflow (even if it's Excel spreadsheets), building a domain glossary, and asking the founder to walk me through the manual version of every process. The goal is to anticipate design requirements before they're articulated.

Week 2: Map the system, not the screens.

Before designing any UI, I map the architecture: What are the core modules? Who are the user roles? What actions does each role take? What data do they need? I create a simple matrix roles × modules × key actions. This prevents the most expensive 0-to-1 mistake: building screens that don't connect into a coherent system.

Week 2-3: Build the design system alongside the product.

Don't wait until you have 50 screens to document components. I start with colour tokens, typography scale, spacing, and 10 core components by the second week. Every component documented early is one less inconsistency fixed later. The system scales with the product instead of being retrofitted after the damage is done.

Week 3+: Start ugly, end polished.

Paper sketches first. Low-fidelity wireframes second. High-fidelity only after the flow is validated with the stakeholder. Jumping to high-fidelity too early is how you spend three days polishing a screen that gets scrapped because the interaction model was wrong.

Always: Design for the second version.

0-to-1 products evolve fast. The feature you ship in month 2 will look different by month 6. I design with scalability in mind flexible layouts, modular components, patterns that accommodate more data or more user roles without a full redesign.

The best skill for 0-to-1 isn't pixel perfection. It's comfort with ambiguity making confident decisions with incomplete information, shipping, learning, and iterating. Every startup needs a designer who can do that.