The Prototype Trap
We used to prototype everything. Figma prototypes for stakeholder buy-in, interactive prototypes for user testing, coded prototypes for developer handoff. Each one took days. Each one was thrown away. We were spending 30% of our project time building things that would never ship.
The question we finally asked ourselves: what if we just... shipped instead?
The Switch
We replaced our prototype phase with what we call "live sketches" - minimal, functional implementations deployed to a staging environment. Not prototypes pretending to be real. Real code, real data, real interactions - just with a very narrow scope.
A live sketch for a checkout flow doesn't have payment processing. But it has real form validation, real responsive behavior, and real loading states. Users can interact with it and we can measure what they do.
The Data Difference
Prototypes give you opinions. Live sketches give you data. We can see where users actually click, how long they spend on each step, where they drop off. This isn't hypothetical feedback from a moderated session - it's real behavior from real people using a real (if minimal) product.
Scoping for Speed
The trick is scope. A live sketch isn't an MVP. It's one flow, one use case, one happy path - deployed and measurable. We scope each live sketch to fit within 3-5 development days. Anything bigger gets split into multiple sketches.
When Prototypes Still Make Sense
We haven't eliminated prototypes entirely. They still work for: testing novel interaction patterns users haven't seen before, stakeholder alignment on visual direction before committing to code, and exploring multiple divergent concepts when the direction is genuinely unclear. For everything else, we ship.