Design Values

Raw
GuidancelatestAnt Design 5.xRetrieved 2026-05-13

Design Values

Source: ant.design/docs/spec/introduce

Ant Design is unique among major design systems in explicitly naming and documenting four core design values that govern its decisions. These values explain why the system is structured the way it is — they are governance principles, not just aspirational statements.


The Four Values

Natural (自然)

Behavior and interaction should feel intuitive and reduce cognitive friction. Ant Design draws from how people naturally perceive and act in the physical world: things that look similar behave similarly; things that look different signal different capabilities; cause and effect are immediate and legible.

In practice: interaction patterns that follow established conventions; animations that mimic physical motion rather than abstract transitions; affordances that match their purpose.

Certain (确定性)

Designers and developers should be able to work with confidence. The system provides clear, consistent rules so there is no ambiguity about how a component should look or behave in a given context.

In practice: predictable component behavior; a complete default state for every component; no edge cases where the "correct" answer is undefined. This value is the primary reason Ant Design is prescriptive — ambiguity slows enterprise development teams.

Meaningful (有意义)

Every design element should serve a clear purpose. Decoration for its own sake is avoided. This is directly connected to the restrained color philosophy: color is used to convey information, signal state, or guide attention — not to create visual interest.

In practice: functional colors (success/error/warning) with consistent semantics; sparse use of brand color; typography hierarchy that communicates importance.

Growing (生长)

The system should be able to expand, adapt, and improve over time without breaking existing products. This means stable APIs, backward-compatible token changes, and a design vocabulary that can accommodate new component types without requiring a full redesign.

In practice: the algorithmic token architecture (seed tokens derive many outputs — a change to one seed propagates consistently); version migration guides; component API stability.


Enterprise Context

Ant Design was created for Ant Financial's internal product ecosystem — complex back-office and financial applications where task completion speed and error prevention matter more than visual expressiveness. This context explains choices that might seem conservative in consumer product design:

  • Color restraint: in a dashboard with 15 data series, competing accent colors create noise, not identity
  • High information density: compact mode, small default control heights, tight spacing defaults
  • Consistency over novelty: familiar patterns reduce training time for power users

The system is now used by Alibaba, Tencent, Baidu, Meituan, and Didi — all at significant scale in enterprise contexts.


Contrast with Other Systems

SystemPrimary design orientation
Ant DesignEnterprise task efficiency; certainty above expressiveness
Material Design 3Expressive, branded consumer products; emotion and delight
CarbonIBM enterprise; data-dense, accessible, neutral
Radix ThemesDeveloper-tool aesthetic; config-first, composable
AtlassianTeam collaboration products; role-based emphasis patterns