Interaction to Next Paint (INP): The 2026 Core Web Vital

INP measures the longest interaction latency observed during a user's visit (technically: the 98th percentile of all interactions, with a few exclusions). An interaction includes mouse clicks, taps on touchscreens, and keyboard key presses. Hover, scroll, and continuous gestures are excluded.

The metric captures three sub-phases: input delay (time from input to event handler start), processing time (time spent running event handlers), and presentation delay (time from handler completion to the next frame painted). All three contribute, and any one of them can blow the budget.

Google's published thresholds are: under 200ms = good, 200–500ms = needs improvement, over 500ms = poor. To pass Core Web Vitals on INP, the 75th percentile of page loads (across the trailing 28 days, in CrUX field data) must be at or under 200ms.

In real-world CrUX data from 2025, roughly 65% of mobile origins pass INP — substantially worse than LCP (around 75% pass) and CLS (around 80% pass). INP is now the most common CWV failure for content sites and the dominant CWV failure for JavaScript-heavy SaaS, ecommerce, and SPA applications.

FID measured only the first interaction. A site that loaded fast and accepted the first click quickly could pass FID even if every subsequent interaction was sluggish. INP closes that loophole: a user clicking through five filters on an ecommerce category page, where each click takes 400ms to respond, will report a poor INP even though the first interaction was fast.

This is why many sites that previously passed FID with comfortable margins now fail INP. The underlying performance was always poor; FID just wasn't measuring it.

In INP audits across roughly 80 production sites, the same five root causes account for the vast majority of poor scores. None of them are exotic — and none of them require a full rewrite to fix.

• Long-running JavaScript on the main thread: heavy event handlers, synchronous analytics calls, expensive React re-renders. Fix by breaking work into smaller tasks with scheduler.yield(), web workers, or requestIdleCallback().
• Third-party tag bloat: chat widgets, analytics, A/B test scripts, consent managers. Each adds main-thread work. Audit, defer non-critical tags, and use a tag manager to control execution order.
• Unoptimised React/Vue/Svelte component trees: re-rendering large component trees on every state change. Use memoisation, virtualised lists, and concurrent rendering (React 18+).
• DOM mutation storms: animations or state changes that force layout recalculation across hundreds of nodes. Batch DOM updates, use CSS transforms instead of layout-affecting properties, and avoid reading layout values during writes.
• Hydration cost on SSR/SSG sites: large JavaScript bundles re-attaching event listeners on hydration block the main thread for hundreds of milliseconds. Use selective hydration, islands architecture (Astro, Qwik), or partial hydration patterns.

Lab tools (Lighthouse, PageSpeed Insights' lab section) cannot measure INP because INP requires real user interactions. Lighthouse reports a Total Blocking Time (TBT) approximation, which is loosely correlated with INP but not the same metric.

The only reliable INP measurement comes from field data: Chrome User Experience Report (CrUX) for population-level data, or a Real User Monitoring (RUM) tool (web-vitals.js, SpeedCurve, Calibre, Vercel Speed Insights) for your own visitors. PageSpeed Insights surfaces CrUX field data — that's the score that affects rankings, not the lab score.

Core Web Vitals are a confirmed Google ranking signal but a small one — Google has described it as a 'tiebreaker' between pages of similar relevance. In our before/after measurements on sites that fixed serious INP problems, organic traffic lift averaged 4–12% over 60–90 days, concentrated on competitive head terms where ranking competition is tight.

The bigger impact is usually on conversion, not rankings. Sluggish interactions correlate strongly with bounce rate and form-abandonment in the data we've reviewed. Fixing INP often produces conversion lift that exceeds the SEO lift by a wide margin — which is the more compelling business case.