Design System As Product
Who is the PM of the design system? (If 'no one', this is a liability not a pro…
$ aoe install @community/provocation-design-system-as-productProjection
Always in _index.xml · the agent never has to ask for this.
DesignSystemAsProduct [provocation] v1.0.0
Loaded when retrieval picks the atom as adjacent / supporting.
DesignSystemAsProduct [provocation] v1.0.0
Label
Is Your Design System a Product or a Liability?
Meta
- License: Apache-2.0
- License Provenance: package-authorship-claim
Trigger
team is building a 'design system' as a side-project, library dump, or token export — without a roadmap, owner, customer, or release process
Question
If your design system has no PM, no release notes, no migration guide, no SLA on bug fixes, no customer interviews with consuming teams, and no metric for adoption — in what sense is it a system rather than a folder of components someone copies and forks? You are not running a design system; you are running a graveyard.
Intent
A design system is a product whose customers are other product teams inside the company. Like any product, it succeeds or fails based on whether its customers adopt it, can rely on it, and have a path to influence it. Most failed design systems were never run as products — they were treated as a one-time export from the design tool, then abandoned when consuming teams forked. Reframing the system as a product (with an owner, roadmap, support channel, deprecation policy, semver) is the only configuration that survives company growth.
Resolution Paths
- Context: Component library in a monorepo, no owner, last release 6+ months ago
- Answer: You don't have a design system, you have legacy. Either staff a small DS team (1 designer + 1–2 engineers + part-time PM) and run it as a product, or formally deprecate it and let teams own their components.
- Context: Tokens exist but no component library; teams copy-paste
- Answer: Tokens are a foundation, not a system. Either accept that your DS is tokens-only (and document that explicitly) or invest in primitives + migration guidance.
- Context: DS owned by design team only, no engineering buy-in
- Answer: Components don't ship from Figma. A DS without an engineering co-owner ships marketing slides, not production UI. Pair design with engineering or accept that consumers will fork.
- Context: DS team exists, has PM, releases regularly, measures adoption
- Answer: You're running it as a product. Continue: solicit customer feedback, maintain a deprecation policy, measure component adoption % and time-to-implement-new-pattern.
Diagnostic Questions
- Who is the PM of the design system? (If 'no one', this is a liability not a product.)
- What is your release cadence? When was the last release?
- How do consuming teams report bugs? Is there an SLA?
- What % of new product surfaces use the DS vs forking? Is this measured?
- Do you have a deprecation policy? When did you last deprecate something?
- Is there a migration guide for breaking changes?
Loaded when retrieval picks the atom as a focal / direct hit.
DesignSystemAsProduct [provocation] v1.0.0
Label
Is Your Design System a Product or a Liability?
Meta
- License: Apache-2.0
- License Provenance: package-authorship-claim
Trigger
team is building a 'design system' as a side-project, library dump, or token export — without a roadmap, owner, customer, or release process
Question
If your design system has no PM, no release notes, no migration guide, no SLA on bug fixes, no customer interviews with consuming teams, and no metric for adoption — in what sense is it a system rather than a folder of components someone copies and forks? You are not running a design system; you are running a graveyard.
Intent
A design system is a product whose customers are other product teams inside the company. Like any product, it succeeds or fails based on whether its customers adopt it, can rely on it, and have a path to influence it. Most failed design systems were never run as products — they were treated as a one-time export from the design tool, then abandoned when consuming teams forked. Reframing the system as a product (with an owner, roadmap, support channel, deprecation policy, semver) is the only configuration that survives company growth.
Resolution Paths
- Context: Component library in a monorepo, no owner, last release 6+ months ago
- Answer: You don't have a design system, you have legacy. Either staff a small DS team (1 designer + 1–2 engineers + part-time PM) and run it as a product, or formally deprecate it and let teams own their components.
- Context: Tokens exist but no component library; teams copy-paste
- Answer: Tokens are a foundation, not a system. Either accept that your DS is tokens-only (and document that explicitly) or invest in primitives + migration guidance.
- Context: DS owned by design team only, no engineering buy-in
- Answer: Components don't ship from Figma. A DS without an engineering co-owner ships marketing slides, not production UI. Pair design with engineering or accept that consumers will fork.
- Context: DS team exists, has PM, releases regularly, measures adoption
- Answer: You're running it as a product. Continue: solicit customer feedback, maintain a deprecation policy, measure component adoption % and time-to-implement-new-pattern.
Diagnostic Questions
- Who is the PM of the design system? (If 'no one', this is a liability not a product.)
- What is your release cadence? When was the last release?
- How do consuming teams report bugs? Is there an SLA?
- What % of new product surfaces use the DS vs forking? Is this measured?
- Do you have a deprecation policy? When did you last deprecate something?
- Is there a migration guide for breaking changes?
Relations
related: [@community/principle-design-philosophy-before-canvas, @community/principle-taste-as-engineering, @community/type-design-token, @community/category-layout-grid]
Source
aoe-engine/examples/frontend-design/primes/compiled/@community/provocation-design-system-as-product/atom.yaml