What is Test Pyramid?

Introduced by Mike Cohn, the Test Pyramid reflects the reality that lower-level tests are cheaper to write, faster to execute, and easier to maintain. Unit tests verify individual functions or classes in isolation, running in milliseconds. Integration tests check that components interact correctly with databases, APIs, or message queues. End-to-end tests simulate real user journeys through a fully deployed stack.

The pyramid shape is prescriptive: most of your coverage should come from unit tests because they provide the fastest feedback loop. As you move up the pyramid, tests become slower, more brittle, and harder to debug when they fail. An inverted pyramid, often called the "ice cream cone anti-pattern," indicates over-reliance on E2E tests that are expensive to maintain.

Modern variations include the "testing trophy" (Kent C. Dodds), which emphasizes integration tests as the sweet spot, and the "testing honeycomb" for microservices, which prioritizes integration over unit tests. The right shape depends on your architecture, but the core principle remains: optimize for fast, reliable feedback.