Case study · 2025 — now · 6 min read

Abadikan. Building the operating system for Indonesian SMEs.

Founding Designer on a product trying to do for small Indonesian businesses what Square, Toast, and Square-for-Restaurants did for the US. Brand, marketing surface, product UI, and the design system underneath — all built from zero in a small room with two co-founders.

Role
Founding Designer (part-time)
Timeline
July 2025 — present
Team
3 founders, 1 designer (me), 2 engineers
Industry
SaaS · SME tooling · Indonesia
Surface
Brand, marketing site, web app, mobile companion
Stack
Figma, Variables, Tokens, Linear, GitHub

Context

Indonesia has roughly 65 million SMEs. Most of them run on a stack of WhatsApp groups, paper notebooks, and a free-tier point-of-sale that no one fully trusts. The founders had spent two years working inside one of these businesses to understand the texture of the problem before they wrote a line of code.

They came to me with a deck and a question: can the back of house feel like Linear instead of like SAP? That question shaped everything that followed.

Approach

Research moves

I sat in three businesses for half a day each — a small bakery in Tangerang, a screen-printing studio in Bandung, and a clinic in central Jakarta. I watched, didn’t prescribe. The pattern that surfaced wasn’t about features; it was about where the day breaks down. Always at the same place: the handoff between the front-of-house staff and the owner who reconciles things at night.

Design moves

The product shipped in three layers. First, a brand that doesn’t pretend to be Stripe — warm, locally legible, in Bahasa Indonesia by default with English as a toggle. Second, a marketing surface that reads like a product, not a brochure: live data, real screenshots, a pricing page that explains itself. Third, the app — built around the nightly reconciliation, not the daily transaction.

Decisions

The hardest call was localization depth. We could ship English-first and translate later, the way most Indonesian SaaS does. Or we could design in Bahasa Indonesia and treat English as the second language. We chose the second. It cost six weeks. It paid back in the first user-test session, where the owner said the product felt like “something a friend made for us.”

Brand system · Logo, type, colour, voice
The brand system, sized for everywhere it has to live — from a 32×32 favicon to a six-foot trade-show banner.
Marketing site · Pricing & product tour
The marketing site reads like the product. Live data, real components, no stock photography.
App home · Nightly reconciliation
The app home is the closing screen, not the opening one. Owners land here to check what happened today, not to start the day.

Outcome

Six months in, the product is in early-access with a small cohort of twenty businesses. The brand and the design system are stable enough that the next two engineers shipped a feature without me in the room. That, more than any number, is the test I run on founding work.

20 Early-access teams
94% Token coverage in shipped UI
6 wk From zero brand to live site
2 Localized markets at launch
Founding design isn’t about doing more. It’s about deciding what not to do, and being able to defend it next month.

Reflection

What I’d do differently next time: start the design system earlier than the brand. We did them in parallel and the brand briefly drove component decisions it shouldn’t have. The fix was small but it cost me a week of refactoring that should never have been on the calendar.

What I’m proud of: the product feels like the founders sound. That isn’t a thing you can fake later. You either bake it in early or you spend years apologizing for the gap.

Next case study The Design Factory — productizing a freelance practice. Or jump to Avocado — self-custody crypto that feels like a banking app.