Web Performance Report 2026
Core Web Vitals pass rates, the speed-to-revenue evidence and the platform shift underneath it — compiled from CrUX, HTTP Archive, Deloitte, Google and web.dev case data, current to July 2026.
From the Headless Web & AI Development toolset.
Every quarter of Chrome UX Report data tells the same story: most of the mobile web still misses Google’s own bar for user experience, and the sites that clear it convert measurably better. The gap between those two groups is one of the most reliably monetizable inefficiencies in digital — large, well documented and directly attackable with engineering. This report compiles the current evidence into a single sourced reference: field pass rates, the controlled speed-to-revenue studies, the platform data underneath them and the replatforming math that follows.
Mid-2026 is a useful vantage point. Interaction to Next Paint has now had two full years as a Core Web Vital, long enough to see which architectures pass it and which fail structurally. Field pass rates have climbed roughly four percentage points a year since 2021, yet the spread between the best and worst platforms has widened to nearly forty points. And the composable stack — headless CMS, React meta-frameworks, edge delivery — has moved from conference-talk material to a line item in mid-market budgets, with the headless commerce market forecast to quadruple by 2032.
Use it three ways: benchmark your own CrUX numbers against the field, borrow the case figures to build an internal business case, and pressure-test the outlook calls against your 2026–2027 roadmap. Every figure carries its source and year; where the strongest studies are older, we say so and note how newer case data has confirmed the direction.
The state of speed
The topline from real Chrome users: 48% of mobile origins and 56% of desktop origins passed all three Core Web Vitals as of the July 2025 CrUX dataset analyzed in the HTTP Archive Web Almanac. That is genuine progress — mobile stood at 32% in 2021, 44% in 2024 — but it still means the majority of the mobile web fails the assessment Google uses to define a good page experience.
The bottleneck is loading speed. On mobile, 77% of origins record a good INP and 81% a good CLS, while only 62% manage a good LCP. The median failing site fails on the fundamentals: slow server responses, oversized hero images and render-blocking resources, the exact problems that platform architecture either solves by default or makes chronic.
Platform choice shows up starkly in the field data. In the November 2025 Core Web Vitals Technology Report snapshot, managed platforms led — Duda at 84.9% of origins passing, Wix at 74.9%, Squarespace at 70.4% — while WordPress, the largest install base on the web, sat at 46.3%. The spread between the best and worst major platforms was 38.6 percentage points, and it has been widening as managed platforms ship performance work their customers inherit for free.
- 48% of mobile origins and 56% of desktop origins pass all three Core Web Vitals. — HTTP Archive Web Almanac, 2025
- LCP is the drag: on mobile, 62% of origins pass LCP versus 77% for INP and 81% for CLS. — HTTP Archive Web Almanac, 2025
- Mobile pass rates have climbed steadily: 32% (2021) → 36% (2023) → 44% (2024) → 48% (2025). — CrUX via HTTP Archive Web Almanac
- Fewer than half of WordPress origins pass Core Web Vitals — 46.3% in the November 2025 snapshot, last among major platforms tracked. — CWV Technology Report via Search Engine Journal, 2025
Share of origins passing all three Core Web Vitals
Platform rows: CrUX / HTTP Archive Core Web Vitals Technology Report, November 2025 snapshot (via Search Engine Journal). Context rows: HTTP Archive Web Almanac 2025, July 2025 CrUX data.
Speed is revenue: the evidence
The canonical study remains Milliseconds Make Millions, run by Deloitte with Google across 37 retail, travel, luxury and lead-gen brands in Europe and the US. Its headline finding: a 0.1-second improvement in mobile site speed lifted retail conversion rates by 8.4% and retail average order value by 9.2%. Travel conversions rose 10.1%. The interval matters as much as the number — one tenth of a second is below conscious perception, and it still moved every step of the funnel.
The bounce side of the curve comes from Google and SOASTA, who built machine-learning models on real-world data from a large sample of mobile ecommerce sites. As load time grows from 1 second to 3 seconds, the probability of bounce rises 32%; at 10 seconds it is up 123%. Google’s earlier DoubleClick research found 53% of mobile visits are abandoned outright when a page takes longer than 3 seconds. These are 2016–2017 studies — still the largest published datasets of their kind — and every major case study since has confirmed the direction and rough magnitude.
The cleanest modern evidence isolates speed as the only variable. Vodafone Italy A/B tested two visually and functionally identical pages where one loaded with a 31% better LCP: the faster page produced 8% more sales. Rakuten 24 invested in passing all three Core Web Vitals and measured a 33.1% higher conversion rate and 53.4% more revenue per visitor. When the experiment design removes every other explanation, the revenue effect persists.
- A 0.1s mobile speed improvement lifted retail conversions 8.4%, retail AOV 9.2% and travel conversions 10.1% across 37 brands. — Deloitte × Google, Milliseconds Make Millions, 2020
- Bounce probability rises 32% as mobile load time grows from 1s to 3s, and 123% by 10s. — Google / SOASTA, 2017
- 53% of mobile site visits are abandoned when a page takes longer than 3 seconds to load. — Google / DoubleClick, 2016
- Vodafone Italy: a 31% LCP improvement, isolated in an A/B test on otherwise identical pages, produced 8% more sales. — web.dev, 2021
- Rakuten 24: passing Core Web Vitals coincided with +33.1% conversion rate and +53.4% revenue per visitor. — web.dev, 2021
Bounce probability by mobile load time (vs 1-second baseline)
Increase in the probability a mobile visitor bounces as page load time grows from a 1-second baseline. Google / SOASTA machine-learning models built on real-world mobile ecommerce data (2017).
The INP era
On March 12, 2024, Interaction to Next Paint replaced First Input Delay as the responsiveness Core Web Vital, and it changed who passes. FID measured only the input delay of the first interaction, a bar nearly every site cleared. INP measures every interaction across the full page lifecycle and grades the whole pipeline — input delay, processing time and presentation delay — at the 75th percentile of real users. Good means 200 milliseconds or less; anything over 500ms is rated poor.
Two years in, INP has become the JavaScript-architecture metric. LCP failures are usually image and server problems; CLS failures are missing dimensions. INP failures come from long main-thread tasks: heavy hydration in client-rendered apps, oversized bundles, and third-party tags — chat widgets, ad scripts, tag managers — competing with user input for the main thread. That is why the failures concentrate on mobile, where 77% of origins record a good INP against a near-universal pass rate on desktop hardware.
Fixing it pays in the same currency as load speed. redBus, the Indian bus-ticketing platform, reworked event handling and state management to cut INP by 72% — and measured a 7% increase in sales. Responsiveness reads to users as reliability, and reliability is what checkout flows monetize.
- INP replaced FID as a Core Web Vital on March 12, 2024. Good is 200ms or less at the 75th percentile; over 500ms is poor. — web.dev / Google, 2024
- INP grades the full interaction pipeline — input delay, processing duration and presentation delay — across every interaction on the page. — web.dev
- 77% of mobile origins record a good INP, making it the second-hardest vital to pass on mobile after LCP. — HTTP Archive Web Almanac, 2025
- redBus improved INP by 72% through main-thread and state-management work and measured a 7% increase in sales. — web.dev, 2023
Core Web Vitals thresholds — and what usually fails them
| Metric | Good (p75) | Needs improvement | Poor | Typical failure causes |
|---|---|---|---|---|
| LCP — Largest Contentful Paint | ≤ 2.5s | 2.5–4.0s | > 4.0s | Slow server response, unoptimized hero media, render-blocking CSS/JS |
| INP — Interaction to Next Paint | ≤ 200ms | 200–500ms | > 500ms | Long JS tasks, heavy hydration, third-party scripts on the main thread |
| CLS — Cumulative Layout Shift | ≤ 0.1 | 0.1–0.25 | > 0.25 | Unsized images and embeds, injected banners, late-loading fonts |
Thresholds are assessed at the 75th percentile of real Chrome users, per metric. Source: web.dev / Google.
The framework landscape
The web’s install base and its developer mindshare now point in different directions. WordPress still powers 41.5% of all websites — 59.2% of sites with a known CMS — per W3Techs as of July 2026, yet it posts the lowest Core Web Vitals pass rate among major platforms. Meanwhile React is used by 83.6% of developers in the State of JavaScript 2025 survey, with Next.js at 59%, making the React meta-framework stack the default choice for new commercial builds.
The commercial layer on top of that stack is growing fast. The headless commerce market reached an estimated $1.74 billion in 2025 and is projected to hit $7.16 billion by 2032, a 22.4% compound annual growth rate. The MACH Alliance’s 2026 research adds a second-order reason to care: 78% of organizations with fully implemented composable (MACH) architecture report clear ROI on their AI investments, against 13% of organizations still in early planning — a six-fold gap that turns architecture into an AI-readiness decision.
The practical read for 2026: your platform increasingly sets your speed ceiling. Server rendering, edge delivery, image pipelines and script isolation are architecture properties before they are optimization projects. Teams on modern stacks inherit passing vitals as the default state, and teams on legacy stacks fund an ongoing remediation program to approximate the same result.
- WordPress powers 41.5% of all websites and 59.2% of sites with a known CMS, as of July 2026. — W3Techs, 2026
- React is used by 83.6% of surveyed developers; Next.js by 59%. — State of JavaScript 2025
- The headless commerce market is projected to grow from $1.74B (2025) to $7.16B (2032), a 22.4% CAGR. — Coherent Market Insights, 2025
- Organizations with fully implemented MACH architecture are six times more likely to report clear AI ROI than early-stage peers (78% vs 13%). — MACH Alliance research, 2026
The 2026 stack in numbers
| Signal | Reading | Source |
|---|---|---|
| WordPress share of all websites | 41.5% | W3Techs, July 2026 |
| WordPress share of known-CMS sites | 59.2% | W3Techs, July 2026 |
| WordPress origins passing Core Web Vitals | 46.3% | CWV Tech Report, Nov 2025 |
| React usage among JS developers | 83.6% | State of JavaScript 2025 |
| Next.js usage among JS developers | 59% | State of JavaScript 2025 |
| Headless commerce market, 2025 | $1.74B | Coherent Market Insights |
| Headless commerce market, 2032 (projected) | $7.16B · 22.4% CAGR | Coherent Market Insights |
The mobile reality
Mobile is where the traffic is and where the performance debt sits. Roughly 62–63% of global web traffic comes from mobile devices per Statcounter, and Adobe Analytics measured mobile at about 56% of US online sales through the 2025 holiday season. Yet mobile conversion rates run at roughly half of desktop’s — around 2% versus around 4% in published ecommerce benchmark indexes.
Part of that gap is structural: smaller screens, checkout friction, and the research-on-mobile, buy-on-desktop pattern. The addressable part is speed. Mobile is exactly where Core Web Vitals pass rates lag (48% versus 56% on desktop), where LCP failures concentrate, and where every study in this report measured its largest effects — Deloitte’s conversion elasticity, the SOASTA bounce curves and the DoubleClick abandonment figure were all mobile studies.
That makes mobile performance the highest-leverage CRO program most teams have available. The traffic is already paid for and the benchmark gap quantifies the headroom; closing even part of the mobile-to-desktop conversion spread compounds across every acquisition channel you run.
- Mobile accounts for roughly 62–63% of global web traffic. — Statcounter Global Stats, 2026
- Mobile carried about 56% of US online sales through the 2025 holiday season. — Adobe Analytics, 2025
- Mobile ecommerce conversion rates run at roughly half of desktop’s — about 2% versus about 4% across published benchmark indexes. — Dynamic Yield ecommerce benchmarks, 2025
- The mobile Core Web Vitals pass rate trails desktop by 8 points (48% vs 56%), and LCP is the widest gap. — HTTP Archive Web Almanac, 2025
The business case for replatforming
Optimization on a legacy platform eventually plateaus. A plugin-laden WordPress theme or an aging monolith can usually be tuned into the "needs improvement" band; getting durably green — and staying green as marketing keeps shipping tags and campaigns — is an architecture property. The published before/after record below is what a step-change looks like when the platform itself moves: Renault measured 14 points lower bounce and 13% higher conversions for every full second of LCP improvement, and Lazada’s 3x LCP gain lifted mobile conversion 16.9%.
The payback math is straightforward to run honestly. Take an illustrative store doing $6M a year online at a 2% conversion rate, currently failing Core Web Vitals with an LCP around 4.5s. A rebuild that lands LCP near 2.0s represents the kind of gain Renault and Lazada published. Apply a deliberately conservative total conversion lift of 10% — well under Rakuten 24’s measured 33% — and the site adds roughly $600K in annual revenue. Against a mid-market headless rebuild that typically lands in the low-to-mid six figures, that pencils to payback inside the first year, with the gain recurring every year after. Treat those inputs as placeholders and rerun the model with your own traffic, AOV and build quotes.
The risk in a replatform is rarely the performance outcome; it is the migration itself. Rankings, redirects and render parity decide whether the new stack inherits the old site’s equity, which is why the discipline lives in URL mapping, redirect strategy and pre-launch parity audits rather than in the framework choice.
- Renault: each 1-second improvement in LCP produced 14 points lower bounce rate and 13% higher conversions, measured across 10M+ visits in 33 countries. — web.dev, 2021
- Rakuten 24: passing Core Web Vitals coincided with +53.4% revenue per visitor and +33.1% conversion rate. — web.dev, 2021
- Lazada: a 3x LCP improvement drove a 16.9% higher mobile conversion rate. — web.dev, 2022
- Agrofy Market: a 70% LCP improvement correlated with 76% less load abandonment. — web.dev, 2022
Published before/after: performance work and business outcomes
| Company | Performance change | Business outcome | Source |
|---|---|---|---|
| Rakuten 24 | Passed all three Core Web Vitals | +33.1% conversion, +53.4% revenue per visitor | web.dev (2021) |
| Renault | Each 1s of LCP improvement | −14 pts bounce, +13% conversions | web.dev (2021) |
| Vodafone (Italy) | LCP improved 31% (A/B tested) | +8% sales | web.dev (2021) |
| redBus | INP improved 72% | +7% sales | web.dev (2023) |
| Lazada | LCP improved 3x | +16.9% mobile conversion | web.dev (2022) |
| Agrofy Market | LCP improved 70% | −76% load abandonment | web.dev (2022) |
All figures as published by the companies via web.dev case studies. Results vary with traffic mix, market and scope of the work.
Our free Site Speed Revenue Calculator applies the studies cited in this report to your traffic, conversion rate and AOV — and the Headless Migration Playbook covers the URL mapping, redirect and render-parity work that protects rankings through a replatform. Both are in the EGGKNITE resource library.
Outlook: five calls for 2026–2027
Mobile Core Web Vitals pass rates cross 50% in 2026 — and the platform spread keeps widening.
Mobile origins passing all three vitals have climbed roughly 4 points a year (32% in 2021 to 48% in 2025, per CrUX via the Web Almanac). The gains concentrate on managed and modern stacks, so the gap between the best platforms (85%) and WordPress (46%) widens even as the average rises.
INP becomes the primary competitive separator among sites that already load fast.
LCP keeps improving industry-wide as CDNs and image pipelines commoditize, while JavaScript payloads and third-party tags keep growing. INP is the vital that punishes that trade directly, and the redBus result (+7% sales from a 72% INP gain) shows the revenue attached to fixing it.
Headless and composable architecture becomes the default for mid-market brands above roughly $10M in online revenue.
A 22.4% CAGR forecast through 2032 (Coherent Market Insights) plus the MACH Alliance’s 2026 finding that composable-mature organizations are six times more likely to see AI ROI. The build cost keeps falling as frameworks and hosting mature, moving the payback threshold steadily down-market.
Performance becomes an AI-visibility factor, extending the speed-revenue link beyond human visitors.
AI crawlers and shopping agents fetch pages at scale, largely without executing client-side JavaScript, and favor fast server-rendered HTML they can parse and cite. As assistant-mediated discovery grows, slow or client-only rendering starts costing citations and agent-driven transactions in addition to human conversions.
The mobile conversion gap narrows first for the fastest sites.
The structural half of the gap (screen size, cross-device journeys) moves slowly; the speed half is directly addressable. Deloitte, Renault and the SOASTA curves all measured their largest elasticities on mobile, so sites reaching sub-2.5s mobile LCP capture disproportionate share of the closing gap.
Sources
- HTTP Archive — Web Almanac 2025, Performance chapter (2025)
- HTTP Archive — Core Web Vitals Technology Report (2026)
- Chrome UX Report (CrUX) — Google (2026)
- Search Engine Journal — Core Web Vitals: open source versus proprietary platforms (CWV Tech Report data) (2025)
- Deloitte Digital × Google — Milliseconds Make Millions (2020)
- Google / SOASTA — Mobile page speed industry benchmarks (Think with Google) (2017)
- Google / DoubleClick — The Need for Mobile Speed (Think with Google) (2016)
- web.dev — Rakuten 24 Core Web Vitals case study (2021)
- web.dev — Renault: measuring and optimizing LCP (2021)
- web.dev — redBus: improving INP increased sales by 7% (2023)
- web.dev — The business impact of Core Web Vitals (Vodafone, Lazada, Agrofy and others) (2022)
- web.dev — Interaction to Next Paint (INP) metric documentation (2024)
- State of JavaScript 2025 — Devographics (2026)
- W3Techs — Usage statistics and market share of WordPress (2026)
- MACH Alliance — Composable foundation and AI ROI research (2026)
- Coherent Market Insights — Headless Commerce Market report (2025)
- Statcounter Global Stats — Desktop vs mobile vs tablet market share (2026)
- Adobe Analytics — Holiday shopping season data (2025)
- Dynamic Yield — Ecommerce benchmarks: conversion rate by device (2025)
Let's grow.
EGGKNITE© 2026 ALL RIGHTS RESERVED — A VERTEX INNOVATION COLLECTIVE COMPANY