Project media

Scaling a Design System Across Multiple Brands

Most design systems are built for one product. We were building one for many.

Company

SYNAPSIS

Year

Q2 2025

Role

Design System Lead Designer


Team

8 people: 3 designers (including me), 2 front-end engineers, 2 mobile engineers, 1 product owner

Key Result

SCALE

Around 64 core components

The system supports three token layers, brand theming and dark mode.

ADOPTION

Used across brands

Nearon and at least 8 apps in the Madhani super app are built on it.

RISK MITIGATION

Licensing risk caught before release

I found that the old system overlapped heavily with the Atlassian Design System. 15 to 20 web components were rebuilt, and almost all the others were adjusted to the new standard.

Background

Our company runs two types of projects in parallel, internal product development and client projects. Our internal product already had a design system called Sensory. Sensory was built for one product, in one context, by one team.

As client projects grew, the team decided to scale Sensory into a multi-brand system that could power both the internal product and client projects, so development gets faster across the board.

My Role

  • Audited Sensory before we started building
  • Proposed the roadmap for the multi-brand system, planned the order of work and got approval from the lead
  • Led the design team through the token, component and property restructure
  • Aligned design and engineering on shared naming and sizing standards
  • Set up the snowflake library for project-specific components

What’s Wrong With The Current System

Before jumping into the build, I ran an audit to understand what we were actually working with. It surfaced five findings.

1. Sensory is not a system

It turned out to be only a Figma component library. There was no shared language between design and engineering.

2. Token structure

Tokens existed as Figma variables, but they were built for a single brand.

Token structure before improvement
Token structure before improvement

3. Component flexibility

Icons, logos and other variable elements were hardcoded into components instead of being replaceable.

4. Component architecture

Some components had missing variants and redundant properties, or were built for one use case instead of being reusable.

5. Properties and naming consistency

There were no shared rules for naming component properties.

Properties and naming before improvement
Properties and naming before improvement

The Risk Nobody Had Caught

During the audit, I found that many Sensory components had significant structural overlap with the Atlassian Design System. The internal product was planned for external distribution and client projects were already running, so continuing to build on top of it wasn't safe. I flagged it and recommended a clean break from all the affected components.

The lead made a different call. The goal was to move faster, so a full rebuild wasn't on the table. The decision was to rebuild the components that looked like Atlassian and keep the more general ones. I disagreed, but I went with the decision. In the end, 15 to 20 web components were rebuilt because of the overlap, and almost all the others were adjusted to the new standard.

Building The Shared Language

The first finding made one thing clear. We needed to build a system from the ground up, and that started with the team.

The engineering team hadn't worked with a proper design system before, so design and engineering had different assumptions about what the system was for. We held regular syncs, wrote a shared naming document and ran workshops together. Engineers reviewed the components on their side and showed us the results, so we could check that the code matched the design.

The goal was to land both sides on the same principle: the design system is one language and one source of truth.

What We Considered

We thought about giving each brand its own separate library. We dropped it because it would be too much to maintain. Every fix would have to be repeated in each library, and the number of libraries would keep growing with every new client. So we went with one shared library, with each brand applied through tokens.

Scaling Up The Design System

The approach was based on atomic design principles, but we extended the model to fit our context.

Design token restructure

We restructured the tokens into three layers: Primitive, Brand and Semantic. The Brand layer is what makes multi-brand possible. Change one value there, and every component that uses it updates automatically.

Token structure after improvement
Token structure after improvement
Brand theming demo

Properties standardization across teams

We made clear agreements between design and engineering. When we say size Medium, we both mean 32px. When we say spacing Large, we both refer to the same token. What gets designed and what gets built should be the same thing.

Component restructure

Before building anything new, we cleaned up what was already there. We merged similar components, completed missing variants and fixed broken properties. Every component had to be reusable, consistent and flexible enough for project teams to adapt without breaking it.

Button component
Button component

Adding flexibility with slot components

Before Figma had native slot support, we built our own. Designers could insert custom content into set parts of a component without detaching it. We didn't want to wait for Figma, because component flexibility was one of the main problems from the audit.

Types of component properties
Types of component properties
Slot component demo

Managing Snowflakes

Working across many client projects means each project will sometimes need things the system doesn't cover yet. Instead of treating those as exceptions, we saw them as a chance to contribute back to the system.

We created a separate snowflake library for these project-specific components. It's shared across projects, and its components get reviewed over time to see if they belong in the main system.

What I'd Do Differently

We never set a baseline, so I can show where the system is used but not how much faster it made the team. Next time I'd measure one thing from the start, like how long it takes to set up a new project.

Next Step: Brand Freedom

We believe branding is more than changing colours and fonts. In the next phase, we want to give the system more freedom to express different brand identities, so it can work for a much wider range of clients.

Better Days Are Yet To Come.

© 2026 Sofyan Hidayat•Built with Fumadocs and Base UI