INP最適化 実践ガイド 2026年版 — 200ms以下のレスポンスを実現する方法
INPはCore Web Vitalsで最も不合格率が高い指標です。scheduler.yield()やWeb Workers、LoAF APIを使ってINPを200ms以下に改善する実践テクニックを、コード例とケーススタディを交えて解説します。
Daniel started in performance work on the SRE side. He spent six years at Spotify on the Web Player team, where he owned the TTI regression budget for the desktop web app and built the internal dashboard that flagged perf regressions per PR before merge. He left in 2023 to join a small consultancy doing performance audits for fintech and travel companies, mostly in the UK and Nigeria. His subspecialty is server-side rendering tradeoffs: when streaming SSR actually helps, when it makes things worse on flaky 4G, and the real numbers behind React Server Components for content-heavy sites. He's a heavy Playwright user for perf testing, mistrusts most npm dependencies on principle, and is currently writing a small Rust tool to diff WebPageTest waterfalls across deploys. Outside of work he coaches a junior dev meetup in Manchester.
INPはCore Web Vitalsで最も不合格率が高い指標です。scheduler.yield()やWeb Workers、LoAF APIを使ってINPを200ms以下に改善する実践テクニックを、コード例とケーススタディを交えて解説します。