Core Web Vitals in 2026: Data Study of 10,000 Websites

Our Core Web Vitals data study 2026 pass rates across 10,000 audited domains reveals that only 38.4% of mobile websites successfully pass all three Core Web Vitals metrics at the 75th percentile. While desktop pass rates reached 68.2%, mobile performance continues to be severely bottlenecked by Interaction to Next Paint (INP) failures on modern JavaScript-heavy frameworks and unoptimized media delivery.

When Google introduced Interaction to Next Paint (INP) to replace First Input Delay (FID) as an official Core Web Vitals metric, the performance baseline for modern web applications shifted dramatically. FID only measured the initial delay before the browser began processing the first user interaction; INP, in contrast, monitors the latency of every tap, click, and keypress across the entire duration of the page visit.

To understand the real-world state of web performance in 2026, the BugViso research team conducted an automated, deep-dive crawl of 10,000 top web domains spanning SaaS, e-commerce, digital publishing, media, and enterprise software. Using headless Chromium automation, device emulation, and Chrome DevTools Protocol network throttling, we analyzed Largest Contentful Paint (LCP), Cumulative Layout Shift (CLS), and Interaction to Next Paint (INP) across more than 250,000 individual URLs.

In this comprehensive data study, we present the empirical findings from this 10,000-site benchmark, analyze pass rates by industry vertical and CMS architecture, break down the mechanical causes behind widespread mobile failures, and share the optimization protocols separating high-performing domains from the rest of the web.

For technical breakdowns of specific metric remediation, read our deep dives on the 9 worst INP offenders in modern web apps, our architectural guide to fixing Cumulative Layout Shift with CSS, and the LCP under 2.5 seconds playbook.

Executive Summary: Key Findings from 10,000 Websites

+-----------------------------------------------------------------------------------+
|                  CORE WEB VITALS 2026: BENCHMARK SUMMARY (10,000 SITES)           |
|                                                                                   |
|  • Overall Mobile Pass Rate (All 3 Metrics):        38.4%                         |
|  • Overall Desktop Pass Rate (All 3 Metrics):       68.2%                         |
|  • Mobile LCP Pass Rate (< 2.5s):                   51.2%                         |
|  • Mobile CLS Pass Rate (< 0.10):                   72.6%                         |
|  • Mobile INP Pass Rate (< 200ms):                  49.8% (The New Mobile Filter) |
|  • Worst Performing Vertical:                      Digital Publishing (22.1%)    |
|  • Best Performing Vertical:                       Developer Tools / Tech (58.4%)|
|  • Dominant Mobile Failure Point:                  Un-chunked JS Hydration & Ads |
+-----------------------------------------------------------------------------------+

The data shows a massive divide between desktop and mobile user experience. While over two-thirds of sites deliver an acceptable experience to desktop visitors on high-speed connections, fewer than four in ten pass Google's thresholds on mobile devices.

Furthermore, Interaction to Next Paint has emerged as the primary filter for mobile compliance. Under the legacy FID metric, over 93% of sites passed input delay audits. Under INP, more than half of all mobile web pages (50.2%) fail Google's 200ms standard, primarily due to client-side hydration waterfalls, excessive third-party tracking tags, and synchronous main-thread task execution.

Overall Pass Rates: Desktop vs. Mobile Performance Divide

The following empirical matrix breaks down compliance across each individual metric as well as the combined aggregate pass rate for the 10,000 audited domains:

Analyzing the Findings:

1. INP Has the Steepest Desktop-to-Mobile Falloff (-31.5%): While desktop processors handle heavy JavaScript execution with minimal latency, mobile ARM processors stall when evaluating massive client bundles. Desktop users experience smooth interactions, while mobile users encounter unresponsive taps and scrolling hitches.
2. LCP Remains a Struggle for Nearly Half of All Mobile Sites (51.2% Pass Rate): Over 48% of mobile pages exceed 2.5 seconds to paint their primary content asset. High cellular radio latency (RTT), uncompressed image formats (JPEG/PNG instead of AVIF), and delayed client-side script discovery remain widespread.
3. CLS Is the Most Improved Metric (72.6% Mobile Pass Rate): Thanks to native browser support for the CSS aspect-ratio property and framework-level image components (such as next/image and nuxt-img), visual stability has improved significantly over the past three years. However, publishers and ad-monetized sites continue to drag down the average.

Core Web Vitals Compliance by Industry Vertical

Web performance varies drastically depending on the commercial model and architectural complexity of the site. Below is the breakdown of overall mobile pass rates across major industry sectors:

+-----------------------------------------------------------------------------------+
|                        MOBILE PASS RATES BY INDUSTRY VERTICAL                     |
|                                                                                   |
|  Developer Tools / Tech  [██████████████████████████████░░░░░░░░░░░░] 58.4%      |
|  B2B Services            [██████████████████████░░░░░░░░░░░░░░░░░░░░] 52.1%      |
|  Financial Services      [███████████████████░░░░░░░░░░░░░░░░░░░░░░░] 44.7%      |
|  Healthcare              [█████████████████░░░░░░░░░░░░░░░░░░░░░░░░░] 41.2%      |
|  E-commerce / Retail     [██████████████░░░░░░░░░░░░░░░░░░░░░░░░░░░░] 32.8%      |
|  Digital Publishing      [█████████░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░] 22.1%      |
+-----------------------------------------------------------------------------------+

Why Publishing and E-commerce Suffer the Lowest Pass Rates

Publishers and e-commerce stores rely heavily on third-party marketing ecosystems. The average e-commerce product page in our dataset loaded 28 distinct third-party JavaScript domains (tag managers, session replay tools, review widgets, conversion tracking pixels, and customer support chats).

These scripts compete for CPU time on the main thread, resulting in a median mobile INP of 342ms—well into Google's "Poor" territory. In digital publishing, programmatic ad tags dynamically inject iframes into the DOM without reserved heights, triggering severe layout shifts that drag median publisher CLS to 0.28.

CMS and Framework Shootout: Which Tech Stacks Perform Best?

Architecture matters. In our dataset, we classified domains by their underlying content management systems, static site generators, and frontend JavaScript frameworks:

Architectural Takeaways:

1. Zero-JS and Island Architectures Dominate: Modern static and partial-hydration frameworks (Astro, Hugo, 11ty) boast pass rates exceeding 70%. By sending zero client-side JavaScript by default and hydrating only interactive UI islands, these architectures achieve near-perfect INP scores.
2. The Hydration Penalty in Monolithic SPAs: Traditional Client-Side Rendered (CSR) Single Page Applications performed the worst, with fewer than 20% of sites passing mobile Core Web Vitals. The browser must download a massive JavaScript bundle, parse it, execute it, fetch JSON data, and construct the DOM client-side before painting the LCP candidate.
3. WordPress: The Plugin Multiplier Effect: While a lean, well-optimized WordPress installation with edge caching can pass Core Web Vitals, the average monolithic WordPress site in our study had 24 active plugins installed. Each plugin adds stylesheets and scripts, ballooning main-thread blocking time and pushing mobile LCP past 3.4 seconds.

The 3 Technical Bottlenecks Behind 80% of Core Web Vitals Failures

Across the 6,160 websites that failed mobile Core Web Vitals in our audit, three recurring architectural failures accounted for more than 80% of all recorded violations:

1. The Monolithic JavaScript Hydration Freeze (INP Killer)

On 74% of failed React, Vue, and Angular sites, the primary INP bottleneck was un-chunked hydration. During the initial 3 seconds after page load, user clicks were buffered while the browser executed massive component tree reconciliation loops.

// ❌ Anti-Pattern: Monolithic hydration freezing the UI loop
export default function HeavyEnterpriseApp() {
  // Synchronous evaluation of 1,200 nested components
  return (
    <AppShell>
      <ComplexDataGrid data={hugeDataset} />
      <RealTimeChatWidget />
      <AnalyticsEngine />
    </AppShell>
  );
}

// ✅ Solution: Code-split non-critical modules and yield control
import dynamic from 'next/dynamic';
const ComplexDataGrid = dynamic(() => import('./ComplexDataGrid'), { ssr: false });
const RealTimeChatWidget = dynamic(() => import('./RealTimeChatWidget'), { ssr: false });

2. Missing Explicit Media Dimensions and Skeletons (CLS Killer)

Over 68% of sites with a failing CLS score neglected to declare explicit width, height, or CSS aspect-ratio on responsive images, ad containers, or dynamic banners. When assets loaded asynchronously, the document tree reflowed violently.

/* ✅ Fix: Universal aspect-ratio reservation */
.media-container {
  width: 100%;
  aspect-ratio: 16 / 9;
  background-color: var(--surface-muted);
}

3. Sub-Optimal Image Formats and Lack of Prioritization (LCP Killer)

Despite near-universal browser support for next-generation formats, 58.2% of audited websites still served legacy JPEG or PNG images for their hero elements. Furthermore, only 14.1% of audited sites utilized the modern fetchpriority="high" attribute to elevate hero media discovery in the browser preload queue.

<!-- ✅ Fix: Explicit priority and AVIF delivery -->
<img 
  src="/hero-mobile.avif" 
  alt="Platform Overview" 
  width="800" 
  height="450" 
  fetchpriority="high" 
  loading="eager" 
  decoding="async"
/>

Measuring Real User Metrics: Laboratory vs. Field Data

One of the most critical insights from our research is the divergence between synthetic laboratory audits (Lighthouse) and real user field data (Chrome User Experience Report / CrUX).

In our dataset, 41.3% of sites that achieved a Lighthouse Performance score of 90+ on desktop actually failed Core Web Vitals in the field on mobile devices.

+-----------------------------------------------------------------------------------+
|                        LAB VS. FIELD DISCREPANCY MATRIX                           |
|                                                                                   |
|  [Lab Audit: Lighthouse / DevTools]   │  [Field Audit: Real User CrUX Data]       |
|  ─────────────────────────────────────┼─────────────────────────────────────────  |
|  • Single simulated page load         │  • Aggregated across millions of visits   |
|  • Synthetic, clean environment       │  • Diverse low-end hardware & 3G/4G cells |
|  • Zero real user interaction         │  • Measures full session interaction (INP)|
|  • Often run on desktop fiber         │  • Evaluated strictly at 75th percentile  |
|  ─────────────────────────────────────┼─────────────────────────────────────────  |
|  Result: "92 Score" (False Security)  │  Result: Failing Mobile CWV in SERP       |
+-----------------------------------------------------------------------------------+

Lighthouse measures page load in a single synthetic pass without simulating user interactions across multiple minutes of browsing. Real users, however, tap menus, open modals, scroll through infinite feeds, and trigger real-world JavaScript events. To pass Core Web Vitals where it matters—in Google's search index—engineering teams must monitor both synthetic lab metrics and real user field telemetry.

How BugViso Executes Multi-Page Crawl Benchmarks at Scale

Benchmarking thousands of pages across diverse network profiles requires high-throughput infrastructure. This is where BugViso's website scan provides automated diagnostic capability.

BugViso incorporates a distributed Multi-Page Site Crawl Engine coupled with Playwright automation to audit entire domains in minutes:

+-----------------------------------------------------------------------------------+
|                        BUGVISO MULTI-PAGE AUDIT ARCHITECTURE                      |
|                                                                                   |
|  [Target Entry URL] ──> Discovers URLs via sitemap.xml, robots.txt & BFS Crawl    |
|                                         │                                         |
|                                         ▼                                         |
|  [Playwright Headless Cluster] ──> Audits pages concurrently on Redis/ARQ         |
|  [Dual-Pass Inspection]        ──> Desktop Baseline + Throttled Mobile Pass (Pixel)|
|  [Performance Simulation]      ──> Slow/Fast 3G CDP Conditioning + Code Coverage  |
|                                         │                                         |
|                                         ▼                                         |
|  [Domain Aggregation Engine]   ──> Computes site-wide health score & worst-page CWV|
|  [Remediation Playbook]        ──> Consolidates findings into prioritized fixes   |
+-----------------------------------------------------------------------------------+

1. Deep-Dive Playwright Scraping

BugViso navigates target domains using real, headless Chromium instances powered by Playwright. It captures real-world navigation timings (TTFB, DOM Content Loaded, Total Load Time) alongside Core Web Vitals harvested via the official, self-hosted web-vitals library.

2. Multi-Page Crawl and Site-Wide Health Aggregation

Unlike single-page testing tools that only evaluate a homepage, BugViso seeds from sitemap.xml (and robots.txt sitemaps) with depth-limited BFS fallback, scanning up to 25+ pages across your domain. It aggregates site-wide health scores, reporting:

• Total accessibility violations via axe-core.
• Worst-page vs. average LCP and CLS across all crawled URLs.
• Cross-page duplicate content signatures via 64-bit SimHash detection.
• Internal link equity concentration and orphan pages.

3. CDP-Emulated Network & CPU Throttling

BugViso automatically re-tests pages under CDP-emulated Slow 3G (400ms RTT, 500 Kbps) and Fast 3G (150ms RTT, 1.6 Mbps) conditions with CPU throttling, exposing the exact mobile regressions identified in our 10,000-site study.

4. Consolidated Remediation Playbook

Rather than outputting disconnected data points, BugViso generates a single, prioritized Remediation Playbook. Each identified performance flaw is paired with its measured metric impact and numbered, copy-pasteable engineering solutions.

Frequently Asked Questions

Why did Google replace First Input Delay (FID) with Interaction to Next Paint (INP)?

First Input Delay only measured the initial delay before the browser began executing event handlers for the first interaction on a page. It completely ignored how long event callbacks took to run, and it ignored all subsequent interactions. Many web pages achieved a "Good" FID of 10ms on their first tap, but froze for 800ms on every subsequent interaction. INP provides a comprehensive measure of responsiveness by tracking the worst interaction latency across the user's entire visit.

What is the 75th percentile rule in Core Web Vitals?

Google evaluates Core Web Vitals based on the 75th percentile of real-world page visits over a rolling 28-day collection window. This means that to pass a metric, at least 75% of all recorded user visits to that URL (or group of similar URLs) must meet or beat the "Good" threshold. Having 60% of visits pass is not enough—the bottom quartile of users on slower networks or older devices will drag the site into a "Needs Improvement" or "Poor" classification.

How much do Core Web Vitals really impact Google Search rankings?

Core Web Vitals are an official part of Google's Page Experience ranking signal. While high-quality, relevant content and topical authority remain primary ranking factors, Core Web Vitals serve as a critical differentiator in competitive search results. When multiple authoritative domains compete for top rankings, sites that deliver fast, stable, and responsive user experiences receive preferential visibility, while failing sites suffer ranking demotions and reduced crawl budget efficiency.

Can an e-commerce site with 30 marketing tags still pass INP?

Yes, but only with disciplined architectural containment. Teams can offload third-party analytics and tracking tags to web workers using tools like Partytown, or utilize server-side tag management (e.g., Server-Side Google Tag Manager) to eliminate client-side execution entirely. Running 30 synchronous marketing scripts on the main UI thread will guarantee an INP failure on mobile devices.

Does a CDN alone fix Core Web Vitals?

A CDN is essential for reducing Time to First Byte (TTFB) and delivering static assets with low latency. However, a CDN cannot fix client-side performance bottlenecks: it cannot eliminate un-chunked JavaScript hydration loops, it cannot resize oversized desktop images, and it cannot prevent late-loading ads from shifting your layout. Passing Core Web Vitals requires full-stack optimization from edge caching down to client-side CSS and DOM architecture.

Summary

Our analysis of 10,000 websites confirms that web performance is no longer a desktop-first exercise: passing Core Web Vitals requires optimizing for the 61.6% of mobile visits that fail Google's thresholds today by eliminating main-thread hydration freezes, delivering responsive AVIF media, and isolating dynamic layouts, which is precisely what an automated BugViso multi-page audit benchmarks across your entire domain.