From 134 to 26 in 12hrs → 10x Faster: Why I Just Nuked Our Design System.

Something felt off.

We weren’t blocked. Nothing was on fire. But we were moving unusually slow.

So I did what I’ve learned to do as a founder:

❌ I stopped asking:

“What are we working on?”

And started asking:

✅ “What’s slowing us down?”

I looked at process.
I looked at ownership.
I looked at engineering.

Then I had a lightbulb moment:

Here’s something I hate to admit:

From Jan 2020 through most of 2023, the first few years of building this company, I didn’t even know what a design system was.

I didn’t know you could use one off the shelf.

I didn’t know how much drag, debt, and decision friction I was creating for everyone around me just by not knowing.

I mean technically I knew “conceptually” but couldn’t understand what needed to be put in process. I was lacking dots to connect backward, concrete consequences to avoid paying…not just following theory from my research / advisors / blogs etc.

So this time, the moment I felt the slowdown, I trusted my gut.

I popped the Figma hood.

And sure enough:

Even though we’d adopted template-only protocol by fall 2023, I’ve taken a passenger seat on design hygiene…and the entropy caught up.

🧱 What I Found

Despite having taken a brutalist mindset — intentionally building a “hostile” product to force behavior validation — and despite already using a template-based system…

…we’d still managed to bloat ourselves with:

• 134 components

• 7 logo variants

• 11 shades of 5 different colors

• 6 heading styles

• 3 body sizes × 3 weights

• Sizing tokens from 3xs to 4xl

• Spacing tokens from 2xs to 4xl

• 5 elevation levels

• 40+ icons (of which we used maybe 5 with the appropriate intentionality to warrant a unique icon for any given use case)

• 9 illustrative icons (so the “cute” icons on top of the “functional” icons)

No one owned the bloat.
No one popped it.

So slowly yet suddenly, every design decision was just-slightly-different.

Uh oh.

Every handoff kept taking increasingly longer than it should should.

It wasn’t just design debt. It was design drag: silent, compounding, expensive.

🔥 So I Nuked It

Twelve straight hours.
No calendar. No Slack. Just Figma and a delete key.

I tore the entire system down to its functional skeleton.

I rebuilt only what the product actually needed:

What could be handed off cleanly — without meetings, checklists, or translation — and wouldn’t sacrifice user experience…that meant minimum modular design system.

✏️ The Rebuild: 26 Components, Fully Tokenized

• Typography: 3 headings x 3 bodies modular for full responsiveness

• Buttons: 2, period.

• Foundational Components: 1 card, 1 table, 1 modal, 1 avatar, 1 dropdown

• Visual Language: 1 interactive response system that covers 99% of our UX patterns
  
  Every remaining component has a purpose.

There’s no more “design debt” because there’s no more design guesswork.

🚀 What That Unlocked

Now, we can:

• Split our design team into modular pods (Creator Portal, Advertiser Dashboard, AI Risk Tool)

• Swat-team designers across surfaces with zero context loss

• Hand off dev-ready files without overhead

• Scale surfaces, states, and compliance logic without slowing down

And here’s how I know it works:

That same project would've taken 2–3 weeks under the old system and still ended in inconsistencies.

Now it takes two days.
And it’s idiot proof.

🧠 What I Know Now

Design systems aren’t about pretty buttons.
They’re not about polish or personality.
They’re not even about speed (though speed is the outcome).

They’re about protection:

• Protecting your team’s time

• Protecting your #1 focus: core value prop of your product

• Protecting against entropy

Even clean systems drift.
Even “good” templates decay without stewardship.

And the second you stop holding the line?
Drag creeps in.

You don’t need more components.
You need fewer decisions.

And sometimes, the only way out is a controlled burn down.

🏁 Final Form

We didn’t simplify to look clean.
We simplified to move cleanly.

From 134 components to 26.
From style preference to operational clarity.
From design bottleneck to cross functional leverage.

If something on your team feels slow, don’t just look at the roadmap.

Look under the design hood.

PS: Normally I wouldn’t do this teardown myself. I’d sit with our team and use it as a shared teardown exercise, a teaching moment.

But we don’t have time for that right now. So I’ll cover it in depth in this week’s learnings meeting instead…using this blog post.

🐦🐦🪨

6:49 am. Wired. But going to bed now.

Kaeya