Cooking a design

I don’t open Figma to “invent from zero.” Usually I already have several tabs open—more than I probably need. One site for spacing, another to see how the navigation is structured, maybe a third because I like the typography even if nothing else fits.

None of them is the thing I’m building. They’re more like raw material. I look at them, break them into parts, and try to assemble something cohesive—a layout that holds together as a system, not just a collection of things I liked.

What I “take”

Over time I realized it’s important to be specific. Not “I like the vibe,” but things like: grid width, heading density, border style, the character of an animation.

If I can’t explain what I’m taking, I’m probably just copying a screenshot. And that usually leads to getting stuck later—you don’t really understand how to adapt it.

References often conflict with each other. One layout is tight and compact, another is built around space and breathing room. I used to try to pick one, but that rarely works.

Now I just start moving things around and see what happens. Gradually the contradictions soften—not because I chose a side, but because I found a balance that works within a single screen.

When it’s “good enough”

There’s a moment when the design stops feeling like a set of sources. It’s hard to define, but you can feel it.

I know I’m close when I can’t point at a block and say, “this is from there.” Not because I hid anything—but because everything follows the same rules.

A consistent rhythm starts to appear: spacing behaves the same way everywhere, the type scale feels aligned, shadows, corners, and details follow shared principles. Everything begins to speak the same language.

If you skip this step, the result almost always looks like a mood board with a URL—visually interesting, but not cohesive.

Why it doesn’t look like the originals

Even if the starting point is someone else’s work, the final result always moves in a different direction.

The reason is simple: my content is different, my constraints are different, and my problems are different. Sometimes there’s more text, sometimes there’s a tight deadline, sometimes things need to be simplified rather than expanded.

That’s where the design starts to separate from the references. It stops being a remix and becomes something of my own—not because I tried to be “original,” but because I was solving a real problem.

I still collect references. I still delete half of what I try.

But that’s not the important part. What matters is what remains—a piece I can stand behind, built from parts I understand and brought to a state where everything inside agrees.