The JavaScript Framework Effect: How React, Angular, Next.js, and Astro Impact Website Quality

"Should I use React or Next.js?" "Is Angular dying?" "Does client-side rendering hurt SEO?" Developers ask these questions constantly. The answers they get are always theoretical — framework maintainers claiming their tool is best, blog posts citing Google's official "we can render JavaScript" position, Stack Overflow threads filled with anecdotes.

Nobody has measured what actually happens when tens of thousands of real websites are built with each framework.

We analyzed 160,000+ websites across 15 JavaScript frameworks, static site generators, and CMS platforms in LLMSE's database and cross-referenced them with SEO, EEAT, WCAG accessibility, readability, and GARM brand safety grades. This is the first public dataset connecting framework choice to multi-dimensional quality outcomes at scale.

The headline finding: your framework doesn't determine your quality — but it strongly predicts it. And the predictions aren't what you'd expect.

The Data

We identified websites running detectable JavaScript frameworks, static site generators, and CMS platforms in LLMSE's database as of March 2026. Detection is based on HTML signatures — CDN imports, framework-specific class names, meta generators, and build tool artifacts.

Framework and Platform Sizes

Category Platform Domains Detection Method
Legacy jQuery 113,019 CDN imports, jQuery global
CMS WordPress 477,394 Generator meta, wp-content paths
CMS Drupal 25,689 Drupal.settings, CSS patterns
Meta-framework Next.js 15,427 __next div, _next/ paths
SSG Jekyll 15,989 Generator meta, GitHub Pages patterns
Builder Shopify 13,942 cdn.shopify.com patterns
Builder Squarespace 12,216 squarespace.com assets
SSG Astro 10,297 astro- prefixes, island attributes
Builder Webflow 9,996 webflow.com assets
SPA React 9,582 react, __reactRoot patterns
SSG Hugo 8,590 Generator meta
CMS Joomla 7,281 Joomla references, /media/system/
SPA Angular 6,691 ng- attributes, Angular patterns
SSG Gatsby 2,013 gatsby- prefixes
Lightweight Alpine.js 1,442 x-data, x-bind attributes
CMS Ghost 865 ghost- patterns

We also detected Remix (399), Stimulus (619), Svelte (183), Preact (175), and HTMX (132), but their sample sizes are too small for reliable grade distributions. We report their data where statistically meaningful.

SEO: The Framework Hierarchy

We define "passing" as SEO grades A, B, or C. The web average pass rate is 2.0%. Framework choice creates a surprisingly wide spread.

SEO Pass Rates by Platform

Rank Platform Graded Pass Rate F-Rate
1 Ghost 452 10.4% 70.4%
2 Astro 6,170 6.7% 86.4%
3 Gatsby 1,139 5.7% 84.3%
4 Jekyll 15,118 4.5% 81.4%
5 Svelte 138 4.3% 89.9%
6 Shopify 1,090 3.9% 91.3%
7 Hugo 6,103 3.9% 85.4%
8 WooCommerce 284 3.9% 89.4%
9 Stimulus 448 3.8% 92.4%
10 Drupal 5,571 3.5% 89.7%
11 Next.js 7,949 3.0% 92.1%
12 WordPress 257,120 2.9% 93.2%
13 Squarespace 7,384 2.5% 92.0%
14 Remix 337 2.1% 94.1%
15 Alpine.js 875 1.7% 94.1%
16 Angular 2,885 1.5% 96.7%
17 Wix 347 1.4% 93.4%
18 Magento 1,432 1.3% 96.2%
19 Webflow 4,022 1.2% 95.1%
20 jQuery 91,791 0.6% 98.0%
21 React 560 0.5% 98.2%
22 Joomla 3,862 0.4% 98.2%
Web average 864,613 2.0% 94.0%

Ghost's 10.4% SEO pass rate is 5x the web average — the highest of any platform in our dataset. Ghost's opinionated architecture (server-rendered HTML, built-in meta tags, automatic structured data, XML sitemaps) makes SEO the default rather than an afterthought.

Static site generators dominate the top tier. Astro (6.7%), Gatsby (5.7%), Jekyll (4.5%), and Hugo (3.9%) all outperform every CMS except Ghost. Pre-rendered HTML with proper meta tags — the SSG default output — is exactly what search engines want.

React (0.5%) and jQuery (0.6%) anchor the bottom. React's client-side rendering generates minimal HTML for crawlers — despite Google's claim that it can render JavaScript, real-world SEO outcomes tell a different story. jQuery's low score reflects its presence on older, unmaintained sites rather than a technical SEO limitation.

Next.js (3.0%) outperforms React (0.5%) by 6x. This is the clearest evidence that server-side rendering and static generation matter for real-world SEO. Next.js and React share the same component model, but Next.js pre-renders HTML that crawlers can parse without JavaScript execution.

The SSR/CSR Divide

Rendering Platforms Avg SEO Pass Rate
Static (SSG) Astro, Gatsby, Jekyll, Hugo 5.2%
Server (SSR) Next.js, Ghost, WordPress, Drupal 4.4%
Client (SPA) React, Angular 1.0%
Legacy jQuery, Joomla 0.5%

The gap between server-rendered (SSG/SSR: 4.8% average) and client-rendered (SPA: 1.0%) is nearly 5x. This isn't a debate anymore — it's data.

EEAT: Who Builds Trust?

EEAT measures the content quality signals that Google evaluates for credibility — expertise signals, author attribution, organizational authority, and trust indicators.

EEAT Pass Rates by Platform

Platform Graded A-Rate Pass Rate
OpenCart 2,341 2.7% 97.1%
Webflow 1,783 22.3% 87.4%
Wix 277 1.4% 77.6%
PrestaShop 129 1.6% 75.2%
Drupal 1,990 12.3% 70.9%
WordPress 189,990 3.8% 69.8%
WooCommerce 199 4.5% 69.8%
Astro 4,195 4.5% 69.3%
Joomla 2,984 3.3% 69.3%
Alpine.js 569 5.8% 65.9%
Shopify 370 17.3% 63.2%
Hugo 5,281 1.5% 62.2%
Jekyll 14,107 0.4% 60.5%
Gatsby 894 5.3% 58.3%
Stimulus 307 8.8% 51.1%
jQuery 79,553 1.7% 39.9%
Ghost 322 0.0% 38.8%
Squarespace 6,124 1.5% 38.1%
Next.js 5,518 2.5% 36.1%
Angular 2,269 1.9% 25.0%
React 378 1.9% 16.1%
Web average 673,667 3.7%

Webflow (87.4% pass, 22.3% A-rate) is the EEAT leader among web development tools. Webflow's template ecosystem emphasizes about pages, team bios, and portfolio sections — exactly the signals EEAT measures.

React (16.1%) and Angular (25.0%) have the worst EEAT scores. SPA frameworks produce application-like interfaces that lack the content signals EEAT looks for — author bios, organizational schemas, expertise indicators. When your "website" is actually a web app, traditional trust signals don't apply.

Ghost's paradox: best SEO (10.4%) but mediocre EEAT (38.8%). Ghost blogs excel at technical SEO (meta tags, structured data, sitemaps) but often lack the multi-page content structures and organizational signals that drive EEAT scores. A well-optimized blog isn't the same as a well-trusted website.

Static site generators cluster mid-range (58-70% EEAT pass). Jekyll's 0.4% A-rate — the lowest among SSGs — reflects its dominance in developer documentation and project pages, which rarely include the author credentials and organizational signals that earn A-grade EEAT.

WCAG Accessibility: The Biggest Surprise

Accessibility is where the framework effect is most dramatic — and most counterintuitive.

WCAG Pass Rates by Platform

Platform Graded Pass Rate A-Rate F-Rate
Ghost 21 85.7% 23.8% 4.8%
React 68 76.5% 50.0% 10.3%
Squarespace 1,012 74.7% 36.7% 18.1%
Jekyll 171 70.2% 34.5% 16.4%
Gatsby 155 67.7% 29.7% 19.4%
Remix 41 65.9% 34.1% 14.6%
Next.js 557 61.6% 28.9% 15.4%
Stimulus 67 61.2% 17.9% 16.4%
Drupal 829 56.8% 23.6% 25.7%
Magento 176 55.1% 44.9% 32.4%
Hugo 305 52.5% 28.9% 25.2%
WordPress 40,257 52.3% 16.8% 30.7%
Webflow 247 47.4% 8.1% 30.4%
Astro 1,087 44.3% 15.4% 40.3%
Angular 310 43.5% 13.5% 35.5%
Joomla 1,696 43.5% 14.3% 34.4%
Shopify 117 41.9% 13.7% 46.2%
Alpine.js 189 39.2% 16.9% 48.1%
jQuery 13,793 37.4% 10.9% 40.7%
Wix 150 11.3% 3.3% 88.0%
OpenCart 194 8.8% 2.1% 78.4%
Web average 122,344 52.7% 18.3%

React has the best WCAG accessibility among JavaScript frameworks — 76.5% pass rate with 50.0% A-grade. This contradicts the widespread belief that SPAs are accessibility nightmares. React's component model, ARIA attribute support, and ecosystem of accessibility-focused libraries (react-aria, Radix UI, Reach UI) produce measurably better accessibility outcomes than most alternatives.

The React paradox: worst SEO (0.5%) and worst EEAT (16.1%) but best accessibility (76.5%). React sites are built by teams that care about component-level quality but don't invest in traditional content signals. The framework attracts developers building applications, not content sites — and applications need accessibility more than SEO.

Wix (11.3% pass, 88.0% F-rate) has the worst accessibility of any platform. Despite being a "no-code" builder marketed to non-developers, Wix's generated HTML consistently fails basic WCAG checks. OpenCart follows at 8.8%.

Next.js (61.6%) significantly outperforms Angular (43.5%) on accessibility, despite both being full-featured frameworks. Next.js's React foundation inherits React's accessibility patterns, while Angular's template syntax and custom elements create more accessibility friction.

Readability: Who Writes Clearly?

Readability A-Rates by Platform

Platform Graded A-Rate Pass Rate F-Rate
Shopify 122 42.6% 82.0% 7.4%
Wix 153 1.3% 81.0% 5.9%
Joomla 1,754 20.0% 76.0% 8.6%
Squarespace 1,071 23.4% 76.3% 11.7%
WordPress 42,591 18.1% 70.7% 10.1%
jQuery 14,371 24.8% 66.8% 18.8%
Alpine.js 198 21.7% 66.7% 22.7%
Jekyll 177 16.9% 64.4% 16.4%
Hugo 325 21.5% 64.3% 20.0%
Astro 1,137 12.7% 63.5% 15.6%
Drupal 870 13.8% 60.8% 13.1%
Gatsby 161 19.3% 59.6% 26.1%
Next.js 579 18.7% 49.2% 39.7%
Angular 327 15.3% 48.0% 39.4%
React 74 9.5% 25.7% 71.6%
Web average 128,682 21.7% 64.4%

React has a 71.6% readability F-rate — the worst of any platform. React sites tend to be data-heavy dashboards and applications with dense technical interfaces rather than readable prose. The Flesch Reading Ease metric penalizes short fragments, code-like content, and technical jargon that's standard in web applications.

Next.js (39.7% F-rate) and Angular (39.4% F-rate) also struggle with readability, reinforcing the pattern: application frameworks produce application content, not readable content.

Content-first platforms win readability. Shopify (42.6% A-rate), jQuery-powered sites (24.8%), and Squarespace (23.4%) produce the most readable content — because they're built for presenting information to consumers, not building interactive interfaces.

The Platform Quality Scorecard

Platform Type SEO Pass EEAT Pass WCAG Pass Read A Domains
Ghost CMS 10.4% 38.8% 85.7% 22.7% 865
Astro SSG 6.7% 69.3% 44.3% 12.7% 10,297
Gatsby SSG 5.7% 58.3% 67.7% 19.3% 2,013
Jekyll SSG 4.5% 60.5% 70.2% 16.9% 15,989
Hugo SSG 3.9% 62.2% 52.5% 21.5% 8,590
Next.js Meta 3.0% 36.1% 61.6% 18.7% 15,427
WordPress CMS 2.9% 69.8% 52.3% 18.1% 477,394
Drupal CMS 3.5% 70.9% 56.8% 13.8% 25,689
Webflow Builder 1.2% 87.4% 47.4% 20.4% 9,996
Angular SPA 1.5% 25.0% 43.5% 15.3% 6,691
React SPA 0.5% 16.1% 76.5% 9.5% 9,582
jQuery Legacy 0.6% 39.9% 37.4% 24.8% 113,019
Web avg 2.0% 48.3% 52.7% 21.7%

The Three Tribes of Web Development

The data reveals three distinct clusters of sites, each with a characteristic quality profile:

1. The Content Web (SSGs + CMS)

Platforms: Ghost, Astro, Gatsby, Jekyll, Hugo, WordPress, Drupal Quality pattern: Strong SEO, strong EEAT, moderate accessibility Who builds here: Content publishers, blogs, documentation sites, marketing sites

These platforms produce server-rendered HTML with structured content. They score well on SEO (meta tags, sitemaps, structured data) and EEAT (author pages, about sections, organizational signals) because their output is designed for consumption by both humans and crawlers.

2. The Application Web (SPAs + Meta-frameworks)

Platforms: React, Angular, Next.js Quality pattern: Weak SEO, weak EEAT, strong accessibility (especially React) Who builds here: Product teams, SaaS companies, web applications

These platforms produce interactive interfaces where the "content" is the application itself. Traditional SEO and EEAT signals don't apply because there's no article to optimize and no author to attribute. But accessibility matters — applications need to work for all users.

3. The Template Web (Builders + Legacy)

Platforms: Webflow, Squarespace, Wix, Shopify, jQuery, Joomla Quality pattern: Variable SEO, variable EEAT, polarized accessibility Who builds here: Small businesses, agencies, non-technical users

These platforms produce whatever their templates and builders allow. Quality depends entirely on the template chosen and the effort invested. Webflow excels on EEAT (87.4%) because its templates emphasize organizational credibility. Wix fails accessibility (88.0% F-rate) because its generated HTML ignores WCAG fundamentals.

The Jekyll-GitHub Pages Pipeline

Jekyll deserves special mention. At 15,989 sites, it's the second-largest SSG in our dataset, and its quality profile is unique:

  • SEO pass rate: 4.5% — above WordPress (2.9%)
  • WCAG pass rate: 70.2% — above the web average (52.7%)
  • 79.7% male-targeted content — the most male-skewing platform in our dataset
  • 62.8% Computer & Electronics category — almost two-thirds are developer documentation and tech project pages
  • 93.0% English — the most linguistically homogeneous major platform

Jekyll sites are overwhelmingly developer documentation and open-source project pages. They score well because developers who choose Jekyll tend to write clean, semantic HTML with proper heading structures and alt text — the same skills that produce good SEO and accessibility.

Infrastructure Patterns

Framework choice predicts infrastructure choice:

Platform Top Server Share #2 Server Share
Next.js Cloudflare 24.4% Vercel 22.7%
Gatsby Cloudflare 18.5% Netlify 16.9%
Astro Cloudflare 30.9% Apache 18.7%
jQuery Cloudflare 34.3% nginx 23.4%
Angular Cloudflare 23.1% nginx 18.7%
React Cloudflare 22.4% nginx 12.0%
Hugo Cloudflare 16.1% nginx 10.3%

Next.js is the only framework where a platform-as-a-service (Vercel, 22.7%) rivals Cloudflare as the top server. This reflects Vercel's role as both the creator and primary host of Next.js. Gatsby shows a similar pattern with Netlify (16.9%), which was Gatsby's primary deployment target before its acquisition.

Astro's Apache presence (18.7%) is surprisingly high for a modern framework — reflecting its adoption in European markets where traditional hosting remains common.

Language and Demographics

English Dominance by Platform

Platform English % Top Non-English
Jekyll 93.0% Chinese (3.6%)
React 83.7% Indonesian (3.5%)
Next.js 83.1% German (2.9%)
Angular 80.1% German (4.8%)
Hugo 76.7% Chinese (6.8%)
jQuery 69.4% Chinese (8.7%)
Astro 60.7% German (11.3%)

Astro has the lowest English share (60.7%) among frameworks — with German (11.3%), Portuguese (7.5%), and Spanish (6.1%) representing significant shares. This suggests Astro has achieved genuine international adoption, not just Silicon Valley traction.

jQuery's Chinese presence (8.7%) reflects its continued dominance in Chinese web development, where jQuery remains the default JavaScript library.

Gender Targeting

Platform Male Female Neutral
Jekyll 79.7% 12.6% 7.6%
Hugo 69.0% 20.1% 10.9%
Gatsby 62.3% 23.4% 14.3%
Angular 59.9% 23.8% 16.3%
jQuery 56.4% 25.4% 18.2%
Astro 56.1% 26.2% 17.7%
Next.js 54.4% 22.5% 23.1%
React 44.8% 41.1% 14.1%

React (44.8% male / 41.1% female) is the most gender-balanced framework — a near-even split that reflects React's broad adoption across consumer applications, social platforms, and e-commerce. Jekyll (79.7% male) and Hugo (69.0% male) skew heavily male, consistent with their developer-documentation profiles.

Key Findings

1. Server-rendered HTML still wins SEO. Static site generators average a 5.2% SEO pass rate — 2.5x the web average. Client-rendered SPAs average 1.0%. The rendering model matters more than the framework.

2. React sites have the best accessibility on the web. At 76.5% WCAG pass rate (50.0% A-grade), React outperforms every other framework and most CMSes. The component model that makes React poor for SEO — encapsulated, reusable elements with explicit props — makes it excellent for accessibility.

3. Ghost is the SEO champion at 10.4% pass rate — 5x the web average. Its opinionated, server-rendered architecture makes good SEO the default output. If SEO is your primary goal and you're building a blog or publication, Ghost is the empirically best choice.

4. jQuery signals technical debt. At 113,019 sites, jQuery is still everywhere — but its 0.6% SEO pass rate, 39.9% EEAT pass rate, and 37.4% WCAG pass rate all trail the web average. jQuery doesn't cause poor quality, but it correlates with older sites that haven't been updated.

5. Next.js is the best all-rounder among JS frameworks. At 3.0% SEO, 36.1% EEAT, 61.6% WCAG, and 18.7% readability, Next.js is the only JavaScript framework that performs above average on both SEO and accessibility. It outperforms React on SEO by 6x while maintaining strong accessibility scores.

6. No-code builders have a quality gap. Webflow excels at EEAT (87.4%) but fails SEO (1.2%). Wix has the worst accessibility of any platform (88.0% F-rate). Squarespace leads accessibility (74.7%) but trails on EEAT (38.1%). The builder you choose pre-determines your quality ceiling.

7. Astro is the rising SSG star. At 10,297 sites with a 6.7% SEO pass rate, Astro already outperforms Gatsby (5.7%) on SEO while maintaining strong EEAT (69.3%). Its 60.7% English share — lowest among frameworks — shows genuine international traction. Worth watching.

Methodology

This analysis covers 160,000+ domains with detected JavaScript frameworks, static site generators, or CMS platforms in LLMSE's database as of March 2026. Detection uses HTML signature matching — CDN patterns, framework-specific attributes, meta generator tags, and build tool artifacts.

A single domain can be detected with multiple technologies (e.g., WordPress + jQuery, Next.js + React). Each domain is counted once per detected platform. Quality grades are computed by LLMSE's analyzers against the HTML content retrieved during classification. The "Graded" column shows the actual sample size for each metric — grading coverage varies because not all domains have been analyzed on all dimensions.

Cross-references were computed using Redis sorted set intersections between app-{Platform} indices and grade indices, sentiment, language, gender, and server indices.

Correlation is not causation. Framework choice correlates with the type of developer, organization, and project — all of which independently affect quality outcomes. A site built on React isn't bad at SEO because of React; it's likely a web application where SEO isn't a design goal.

All data is aggregated — no individual site is identified. You can analyze your own site's quality scores at llmse.ai.


This analysis was conducted using LLMSE, which has classified over 1.5 million websites across SEO, EEAT, WCAG accessibility, readability, and GARM brand safety dimensions. All data reflects the database as of March 2026. To analyze your own site, visit llmse.ai/classify.