Mint V8 Design System
Design System Lead · Clear (ClearTax) · 2023
The claim.
How I tested it.
What I'd do differently now.
The claim.
What this project actually taught me about design systems.
Adoption is a design problem.
Not an org problem. Not a mandate problem. A design problem.
The chaos we inherited.
Every team had built their own.
Four teams shipped four action dropdowns. They all work, they're all slightly different, and now we maintain all of them.
Priya, Senior Product Designer · Clear
4 teams · 4 action dropdowns · 4× maintenance burden
Years of every team solving the same problem in isolation.
Three symptoms. One problem.
Same patterns shipped multiple times across teams. Filter panels, dropdowns, tags — every one of them re-invented.
No central team responsible for the canonical components. Each crew solved the same problem in isolation.
Each new theme would require forking every component. Two suites needed different brands but the system couldn’t express variation.
Tokens before components.
Even at the cost of shipping faster.
Atoms → molecules → organisms.
And the rule that was missing.
PCS Logic.
Prefix · Content · Suffix.
A shell. Three slots. Compose, don't invent.
Adoption is a design problem.
Back to the claim.
Mandates create resistance.
Especially for teams who'd just built their own filter panel.
Four mechanisms instead of one mandate.
- 01 Weekly PCS workshops. 45-minute sessions where designers across teams built components together.
- 02 Shared Figma playground. Interactive PCS components for live experimentation — composing felt faster than inventing.
- 03 Token sync pipeline. Tokens exported to JSON for engineers — no design-to-code translation step.
- 04 Gradual deprecation. 7–8 screens migrated per sprint, never all at once. Adoption compounded without resistance.
I was skeptical. Once I built with PCS, I couldn't go back.
Vikram, Senior Product Designer · Clear
Every team chose it themselves.
Make the central library easier than maintaining your own.
Every team will choose it themselves.
A design system isn't a Figma file.
It's a team.
What I'd do differently.
- 01 Build the central team before the library. I started with tokens, then components, then adoption — then formed the team. Reverse it. Form the team in week one; let the library grow inside it.
- 02 Instrument adoption from day one. The four mechanisms worked, but I read the room by feel — week by week, in person. "Composing felt faster than inventing" deserved a number. Next time I want time-to-compose, variant velocity per team, and workshop-to-shipped lag running before the rollout starts.
- 03 Design the conditions for the PCS moment, not the moment itself. PCS surfaced in a working session in week six. I can't codify the discovery, but I can codify the conditions — weekly cross-team sketching sessions from week one, with the right mix of designers in the room. The pattern would still take time to emerge, but it would emerge on purpose.
Thank you.
ashishkhoshya.com · linkedin.com/in/ashishkhoshya