Jest Mocking Patterns That Save Time

Jest's mocking is genuinely powerful — and a place where teams burn weeks chasing flaky tests. The patterns below have produced reliable tests for me, in roughly the order I reach for them.

jest.mock() for module-level swaps

When a module has a heavy dependency (database, HTTP client, third-party SDK), jest.mock() at the top of the test file replaces the module with stubs. This is the most common mock and rarely the source of bugs — it is hoisted, applies once, and is easy to reason about.

jest.fn() for inline call tracking

When you want to assert that a callback was called, jest.fn() is what you pass. expect(callback).toHaveBeenCalledWith(args) verifies what happened without needing module-level mocking.

Avoid spying on production code

jest.spyOn(obj, 'method') replaces the method on the original object. If you forget to restore it, the next test sees a corrupted object. Use spies in tight scopes and call mockRestore() in afterEach. Better: refactor so the dependency comes in via a parameter and you can pass a mock directly without touching the original.

Reset mocks between tests

Set clearMocks: true (or restoreMocks: true) in jest.config so call counts and implementations reset before each test. Forgetting this leaks state across tests in ways that are maddening to debug.

When mocks lie to you

Heavy mocking can produce green tests that hide real bugs. If your test mocks a complex piece of behavior, the test is now mostly verifying the mock — not your code. For those cases, write an integration test against the real dependency (in a sandbox or test container) and skip the unit test.