Google web.dev: performance and Core Web Vitals

Evidence class: platform measurement guidance, not an accessibility standard or a revenue experiment. Read: 2026-09-13.

Verified thresholds

Good Core Web Vitals are Largest Contentful Paint (LCP) at most 2.5 seconds, Interaction to Next Paint (INP) at most 200 milliseconds, and Cumulative Layout Shift (CLS) at most 0.1. Assess the 75th percentile and separate mobile from desktop. These measure loading, responsiveness, and visual stability, respectively. Source: Web Vitals, read 2026-09-13.

Responsiveness is not completion

INP measures the wait until the browser can present the next frame after relevant input. It does not measure all asynchronous network work or its eventual result. A prompt loading state can coexist with slow completion. Source: Interaction to Next Paint, read 2026-09-13.

Website application

Our proposed acceptance checks combine field metrics with task-specific measurements: useful search result, durable save, completed upload, or confirmed submission. Inspect slower devices and failure paths. Preserve space for loading content to avoid moving the intended target. Never label a pending effect as completed merely to appear fast.

Limits

A good technical score is not proof of understandable navigation or successful work. Tognazzini's 50 ms acknowledgment advice (Tognazzini) and the popular 400 ms rule (Myths) concern different interactions and measurement contexts. Do not turn any single number into a universal completion budget. Recheck the thresholds above at each review of this note: web.dev says stable Core Web Vitals change at most once a year, and INP replaced First Input Delay as a stable metric in 2024. Source: Web Vitals, read 2026-10-29.