---
title: "Camelot Design System — Maximilian Dahl"
description: "Product design case study by Maximilian Dahl: Camelot, the design and component system of car2go and SHARE NOW, layered from brand to interface, for iOS, Android and the web."
url: https://max.does.systems/en/work/camelot/
lang: en
---

Product Design Case Study · Design Systems · Merger: car2go → SHARE NOW · car sharing · 2019 –⁠ 2021

# A system built to survive a rebrand

car2go built its app, web and B2B tools in separate teams. I built and led Camelot, with the brand as its bottom layer. That carried the rebrand to SHARE NOW. When we rebuilt the web, I reorganised it by purpose, so app and web stayed one.

Read the case · RJump to results

Role in the project · Built and led · Product designer, leading the founding group from 2019

Platforms · iOS, Android and web · plus B2B tools, websites and print

Responsibility · Architecture and rules · Layers, naming scheme, principles and icon library; with brand, marketing and engineering

Status · Shipped · +2 · As of September 2026 · Shipped · Carried on after the rebrand as SHARE NOW Design System 2.0 · Continued · After I left in 2021, by a design colleague in Stuttgart · Open · Impact figures are not documented · and 2 more states

As of September 2026

**Shipped · +2**

As of September 2026

- Shipped
   Carried on after the rebrand as SHARE NOW Design System 2.0

- Continued
   After I left in 2021, by a design colleague in Stuttgart

- Open
   Impact figures are not documented

I did not build a system for one brand. I put the brand underneath the system so it could carry a change.

Camelot Design System

## At a glance

- 4 layers
   from the brand up to the interface
   **4 layers**from the brand up to the interface · Scope of the system, not a measured outcome

- 2,239
   components in 68 sets, one naming scheme
   **2,239**components in 68 sets, one naming scheme · Size of the design library

- 3 platforms
   iOS, Android and web from one system
   **3 platforms**iOS, Android and web from one system · Plus B2B tools, websites and print

1. Problem
   Organised by purpose
   Four layers from brand to interface instead of one library per platform.

2. Intervention
   The brand at the bottom
   When the brand changes, the change runs up through every layer.

3. Outcome
   Carried the rebrand
   Camelot became SHARE NOW Design System 2.0.

Role in the project**Built and led**Product designer, leading the founding group from 2019

Platforms**iOS, Android and web**plus B2B tools, websites and print

Responsibility**Architecture and rules**Layers, naming scheme, principles and icon library; with brand, marketing and engineering

02 · Starting point

## Every team built its own components

At car2go, teams and products grew fast: the customer app, webviews, B2B, registration, later SHARE NOW Pass. A shared system existed only on the design side; engineering teams had no common source for components, patterns or templates and built much of it themselves.

What followed

1. 01
   Duplicate code
   The same building blocks in several codebases.

2. 02
   Slow delivery
   Every feature started from scratch.

3. 03
   Inconsistent brand
   Users experienced every product a little differently.

4. 04
   No single source of truth
   Nobody knew which component was the real one.

Before: one library per platform

1. App styleguide
   Customer app with its own components.

2. Web styleguide
   Website and webviews, maintained separately.

3. B2B styleguide
   Tools for business customers.

4. BA styleguide
   Yet another world of its own.

Splitting by platform made ownership easy and collaboration hard. Every team worked in its own silo, new components were hard to contribute, and we kept rebuilding the same building blocks.

The Camelot scope named three job stories, and the product teams add a fourth. What they all share: nobody should have to build the same building blocks twice:

- **Company**Job story
   When · new products and teams come in,
   I want · one consistent experience across all products,
   so · users reach their goals the easiest way.

- **Design**Job story
   When · I design a solution for a new feature,
   I want · a scalable design system,
   so · I can build solutions and experiences faster.

- **Engineering**Job story
   When · I build a feature,
   I want · a scalable component system,
   so · I don't have to rewrite every component each time.

- **Product teams**Job story
   When · I build a feature for the app, a webview or the web,
   I want · to use the same building blocks,
   so · the brand looks the same everywhere, even when it changes.

The first three are paraphrased from the Camelot scope, where they were written as user stories; the situation comes from its problem statement. The fourth sums them up for the product teams.

Context

The decision

One system for every platform, not a design-only UI kit. In 2019 we formed a founding group for it, prompted by the new Spotify-style team structure, and I led it. The price: a system for everyone needs everyone's agreement, from brand and marketing to every engineering team.

03 · Layers

## Organised by purpose, not by platform

The order of the layers is the real decision. The brand sits at the very bottom, then the foundations, then the interactive UI and the public-facing pages.

A big influence was Brad Frost and his Atomic Design: small building blocks that combine into larger ones.

Four layers · from the base up

1. Brand
   Colours, illustrations, vehicles and imagery as the foundation.

2. Foundations
   Palette, typography, grid, breakpoints, buttons, input fields.

3. Interactive UI
   For devices and desktop: map pins, sheets, navigation, tables, forms.

4. Public-facing UI
   Website and landing pages, responsive, built from the same foundations.

Context

The decision

The brand sits at the bottom, not on top. When it changes, the change runs up through every layer instead of being patched component by component. The price: every change to a colour or a typeface takes effect everywhere at once and has to be agreed with engineering.

Foundation: the pill buttons · Live in the lab

max.lab · Camelot · Live · rebuilt · rebuilt · Rewritten for this site, not the original code. Where a detail is approximated, the docs say so.[Docs](https://max.does.systems/lab/#/camelot/komponenten/button?lang=en)

Four styles, three heights from 40 to 64 pixels, eight states, built from the foundation tokens.

The names follow the same order: category, subcategory, type, component. “Interactive UI / Device / Messaging / ToastBar” tells you where a building block lives and what it is for before you open it.

This order came later. When we rebuilt the website and the web products for private and business customers, web app elements also found their way into the native apps. There they could not feel different from everything around them, or they would have confused people. So I developed the system further and proposed this order as the new structure in April 2021: for consistency, for flexibility and as the link between app and web. Deliberately not as a hard switch, but as a transition out of the existing libraries.

04 · Platforms

## Native in shape, consistent in behaviour

Because the foundations are shared, platforms only differ where the platform demands it. The toggle follows one form on iOS and the web and another on Android; colours, states and spacing stay the same. Typography comes in dedicated sets for web, iOS and Android.

The same toggle, true to each platform · Live in the lab

max.lab · Camelot · Live · rebuilt · rebuilt · Rewritten for this site, not the original code. Where a detail is approximated, the docs say so.[Docs](https://max.does.systems/lab/#/camelot/komponenten/toggle?lang=en)

Switch the platform and compare: the shape follows the platform, colour and states follow the system.

A whole screen, two platforms · Live in the lab

max.lab · Camelot · Live · rebuilt · rebuilt · Rewritten for this site, not the original code. Where a detail is approximated, the docs say so.[Docs](https://max.does.systems/lab/#/camelot/komponenten/pattern-booking?lang=en)

The booking screen from the same parts. Switch between iOS and Android, selection and confirmation: the shape changes, the behaviour stays.

The library grew to 68 component sets and 2,239 components, all following one naming scheme, with a dedicated area for native iOS patterns.

05 · Rules for change

## The library mirrors the code

A system only stays shared if everyone knows how to extend it. New components had to follow six principles, and every change to a typeface, a colour or an existing component had to be agreed with engineering, because it takes effect everywhere at once.

Principles · Six principles for new components

1. Search first
   Check the library first. The solution probably exists already.

2. Scalability
   Don't build just for the task at hand; build so others can reuse it.

3. Consistency
   Don't reinvent the wheel. Bring it in and make it better.

4. Snowflakes
   Avoid one-off components, or have the facts to justify them.

5. Styleguide
   Use the defined palettes and type; agree anything new with brand, marketing and design.

6. Collaboration
   The library mirrors the components in code. Changes only happen together with engineering.

A justified one-off · Live in the lab

max.lab · Camelot · Live · rebuilt · rebuilt · Rewritten for this site, not the original code. Where a detail is approximated, the docs say so.[Docs](https://max.does.systems/lab/#/camelot/komponenten/car-pin?lang=en)

The map pin is the domain-specific building block: only for the map, but made of the same tokens and with clear states.

The pin where it lives · Live in the lab

max.lab · Camelot · Live · rebuilt · rebuilt · Rewritten for this site, not the original code. Where a detail is approximated, the docs say so.[Docs](https://max.does.systems/lab/#/camelot/komponenten/pattern-map?lang=en)

Map, pins and reservation sheet together. Collapse and expand the sheet and switch the platform.

Later, Storybook became the documentation in code, synced automatically into the native apps. We also had our own icon library and worked with marketing to carry the system all the way into print. In 2021 the library moved to Figma.

### Who decides

At Camelot, the design team made the decisions, and made them together. The final sign-off was mine, as gatekeeper. The developers came to me with systems of their own, since iOS was built in Swift and Android natively. That became one shared route: we kept developing Camelot together, tested it pixel-perfect and connected the platforms through the global foundations.

From proposal into the apps

1. Together in the design team
   The design team decides; the work is done together

2. Sign-off
   Me as gatekeeper, at the end of the review

3. Native build
   The developers build in their own systems, in Swift and for Android

4. Pixel-perfect test
   Checked together against the design, connected through the global foundations

My part

I built and led Camelot: the founding group, the layered order, the principles, the icon library and the proposal for the new structure. Design and engineering put it into practice together.

### Accessibility

Accessibility was a focus of the customer-app work at car2go and SHARE NOW. In Camelot's six principles, though, it was not a rule of its own. The rebuild now measures against WCAG: white on the primary action blue reaches only about 2.9:1, less than the 4.5:1 that AA requires for normal text.

Learning

Libraries split by platform make ownership comfortable and the system lonely.

06 · The rebrand

## Camelot became SHARE NOW Design System 2.0

The move from car2go to SHARE NOW was the stress test for the bottom layer: a new brand met a system that had deliberately put the brand underneath everything else. Camelot shipped and carried on as SHARE NOW Design System 2.0. After I left, a design colleague in Stuttgart took it forward.

The decision came with the 2019 merger: the app had to be rebuilt so it looked like neither car2go nor DriveNow, and the entire design library was to be rebranded. That is where the decision came from to put the brand in its own layer underneath the system. The order by purpose only followed later, when we rebuilt the web products.

Learning

During the merger we rushed into a single feature without assumptions, information architecture or agreement with product, and threw it away months later. Since then I start with the roadmap and what already exists: first clarify what's there and who needs it, then build.

See also:[Compass Design System](https://max.does.systems/en/work/compass-design-system/)

07 · Outcome

## What came of it

We never measured the goals from the scope. What is documented: Camelot shipped, carried the rebrand and was continued after I left.

1. In daily work
   Storybook as documentation in code, synced into the native apps
   Part of how design and engineering work

2. Rebrand
   Carried on as SHARE NOW Design System 2.0
   Rebrand the entire design library

3. Library
   68 component sets, 2,239 components, one naming scheme
   One source for every platform

4. Platforms
   Native shapes and type sets per platform, the same colours and states
   iOS, Android and web from the same foundations

5. In the guild · open
   not documented
   Part of the frontend guild, with at least one vocal advocate

6. At scale · open
   not measured
   Replace duplicated foundation components in 80% of teams

08 · Looking back

## What I learned and would do differently today

### Caveats

I built and led Camelot; design and engineering implemented it together, and after I left in 2021 a design colleague in Stuttgart took it forward. What is documented is its structure, scope and course; there are no figures on its impact. As of April 2021.

### Learnings

1. The brand belongs at the bottom
   A system survives a rebrand when the brand is a layer of its own, not something baked into every component.

2. Organising by purpose means spreading ownership
   Libraries split by platform are comfortable but lonely. Organised by purpose, teams work on the same building block instead of copies.

3. Clarify what exists, then build
   During the merger we rushed into one feature and threw it away months later. Since then I start with the roadmap and what already exists.

4. A system only stays shared if people can extend it
   Clear principles for new components mattered more than any single component.

### What I would do differently today

1. Make the goals measurable
   We had clear success criteria but no measurement. Today I would count the duplicated foundation components before starting, so a goal like 80% of teams can actually be proven.

2. Spread ownership earlier
   Ordering by purpose and shared ownership only came with the 2021 proposal. Today I would decide from the start who maintains which layer.

3. Accessibility as a principle of its own
   Contrast, focus and keyboard to WCAG AA belong next to the six as a seventh rule, with values in the tokens instead of in review.

> A design system doesn't prove itself at its first release. It proves itself at the first rebrand.
> 
> Maximilian Dahl

fin (for now)

**Case notes · How to read this case.**

Evidence · The Camelot Figma library with its pages on structure, principles and history, plus my own account. Impact figures are not documented and so are not shown here.

Rebuild · The embedded components are rebuilds from max.lab, derived from the library's styles and variants. Like the app, they are set in the platform typefaces SF Pro and Roboto; brand assets, photos, vehicle renders and the brand typeface are not included.

Name · Camelot 2.0 is the SHARE NOW Design System 2.0.
