The global Core Web Vitals pass rate just dropped: eCommerce stores are paying the price in conversions

The global Core Web Vitals pass rate has declined for the first time in over two years, driven largely by eCommerce sites adding more third-party scripts and dynamic content. If your store's mobile performance is degrading, audit your script inventory, compare your Lighthouse scores against real-user field data in Google Search Console, and set up continuous monitoring to catch regressions before they cost you conversions.

For the first time in over two years, the trajectory for web performance is pointing downward. The global Core Web Vitals pass rate, the share of websites meeting Google's thresholds for Largest Contentful Paint, Cumulative Layout Shift, and Interaction to Next Paint, has begun to decline, reversing a steady climb that had lasted since the metrics were introduced as ranking signals.

The cause is not a mystery. The web is getting heavier. And eCommerce sites, with their growing stack of third-party scripts, dynamic content, and complex checkout flows, are leading the regression.

The script tax eCommerce stores are paying

According to industry benchmarks published in 2026, eCommerce sites in script-heavy verticals now show average mobile LCP scores of 3.7 to 4.3 seconds, well above Google's 2.5-second "good" threshold. The average web page loads 20 external scripts totalling around 449KB, but eCommerce pages often far exceed that, with consent management platforms, personalisation engines, A/B testing tools, analytics suites, chat widgets, and payment integrations all competing for the browser's main thread.

Each individual script may be modest. Collectively, they are devastating.

And the cost is not abstract. Shopify's enterprise data shows that for every 100 milliseconds slower a store loads, conversion tends to be about 3.5% lower. Shopify also found that a 0.1-second improvement in site speed leads to an 8–10% increase in conversion rates for retail. These are not projections; they are measurements from real stores serving real customers.

Scale that to your revenue. A mid-market store generating £5 million annually that slows by just 300 milliseconds, less than the blink of an eye, is looking at a conversion drop of roughly 10%, or £500,000 in lost sales.

When slow is as dangerous as down

Performance degradation has traditionally been treated as a lower priority than outages. If the site is up, it is working. If it is working, performance is a "nice to have."

That attitude is increasingly expensive. According to LogicMonitor's research, 65% of eCommerce leaders now report that slow performance is as damaging to their business as downtime, with 51% having experienced over $1 million in negative economic impact from performance issues.

The difference between an outage and a performance regression is visibility. An outage triggers alarms, mobilises teams, and gets fixed in hours. A performance regression, say, a new analytics script that adds 400 milliseconds to every page load, can persist for weeks before anyone notices. It does not break anything. It simply makes everything slightly worse. Conversion rates drift down. Bounce rates drift up. Revenue erodes slowly, and the cause is invisible in the metrics most teams monitor.

Why synthetic scores lie to eCommerce teams

Most eCommerce teams monitor performance by running Lighthouse or PageSpeed Insights on a handful of key pages, checking the scores, and moving on. If the numbers are green, performance is "fine."

Synthetic tests run in controlled environments with consistent hardware, consistent network conditions, and a single geographic location. They do not run third-party scripts the way real browsers do. They do not account for the consent banner that loads before the page content, the personalisation engine that injects recommendations after first paint, or the inventory API that returns slowly during peak traffic.

Real users experience your site differently. A returning customer in regional Australia on a 4G connection with a three-year-old Android device encounters a fundamentally different page than the one Lighthouse tested in a data centre. And it is that customer's experience, not the synthetic test's score, that determines whether they convert.

This is the core problem. Lab scores are improving. Field data is declining. The gap between what you measure and what your customers experience is growing.

The eCommerce-specific performance challenge

eCommerce sites face performance problems that most other websites do not:

1. Product pages are template-driven but data-variable. A single product template might perform well for a product with three images and no variants, and fail CWV for a product with twelve images, six colour options, and a size chart loaded from a third-party service. Testing one product page tells you nothing about the ten thousand others.

2. Checkout pages accumulate scripts. Payment gateways, fraud detection, address validation, tax calculation, loyalty point integration, the checkout is where the most JavaScript lives, and where performance matters most. Yet it is often the least-tested page because synthetic tools cannot easily navigate through a multi-step checkout flow.

3. Seasonal traffic creates transient performance problems. A site that passes CWV in July may fail in November when traffic triples and CDN edge servers are under pressure. These regressions are invisible to periodic testing.

4. Marketing campaigns introduce performance volatility. A new retargeting pixel, a promotional banner with unoptimised assets, or a flash sale widget loaded from a third-party CDN can degrade performance for exactly the traffic the campaign is driving. The irony is acute: the marketing spend brings visitors, and the marketing technology drives them away.

From periodic testing to continuous monitoring

The declining CWV pass rate is not a problem that can be solved with a quarterly performance audit. By the time you audit, the regression has already cost you conversions, SEO visibility, and revenue.

What eCommerce teams need is continuous real-user monitoring that tracks Core Web Vitals across every page template, every device class, and every geographic region, not as a periodic check, but as a live operational metric alongside revenue and conversion rate.

AuditIQ eCommerce monitoring approaches this as part of comprehensive storefront monitoring. When a deployment introduces a performance regression, when a third-party script update degrades interactivity, or when a specific product page template fails LCP for mobile users, the problem surfaces immediately, not in the next scheduled audit.

What eCommerce teams should do now

1. Compare your lab scores to your field data. Google Search Console's Core Web Vitals report shows how real users experience your site. If your Lighthouse score is 90 but your field LCP is 4.1 seconds, you have a measurement problem, not a performance problem.

2. Audit your third-party script inventory. Count every script on your product pages, category pages, and checkout. Twenty scripts is the average; if you are above that, you likely have redundant or unused integrations that are costing you performance and revenue.

3. Segment performance by page type, device, and region. Aggregate CWV scores hide the pages that are failing. A 3.7-second mobile LCP on product pages can be masked by fast performance on your static about page.

4. Monitor continuously, not periodically. A CWV regression that lasts two weeks before your next audit costs you conversions every single day. Continuous monitoring catches regressions within hours.

5. Treat every new script as a performance risk. Before adding a new third-party integration, measure its impact on LCP, CLS, and INP in staging. After deployment, verify the impact in production with real-user data.

The mobile web is getting slower. eCommerce sites are getting heavier. And the gap between synthetic lab scores and real-user experience is widening. The stores that close that gap, with continuous monitoring, disciplined script management, and real-user data, will convert more visitors. The stores that do not will wonder why their traffic is growing, but their revenue is not.

AuditIQ monitors your storefront's real-world performance alongside checkout integrity, SEO health, and analytics accuracy, giving eCommerce teams a single view of the metrics that drive revenue.

Start your free AuditIQ trial and see how your store's real-user performance compares to what your synthetic tests are reporting.

Others also read:

• From 39% to 55.3%: Four years of Core Web Vitals progress
• The 100-milisecond rule of Core Web Vitals
• Cumulative Layout Shift (CLS): What is it, its impact and how to control

FAQs

1. Why is the global Core Web Vitals pass rate declining if Google has been pushing performance improvements?

The decline is driven by web complexity growing faster than optimisation efforts. More third-party scripts, heavier JavaScript frameworks, and dynamic content are being added to sites faster than performance improvements at the platform level can offset them. eCommerce sites are particularly affected because they tend to load more scripts than other site types.

2. My Lighthouse score is green. Why should I be concerned about my real-user CWV data?

Lighthouse tests run in a controlled environment without the third-party scripts, consent banners, and personalisation engines that real browsers load. Your Lighthouse score can be 90 while your field LCP is 4+ seconds for mobile users on slower connections. Google Search Console's Core Web Vitals report shows actual field data from real visitors, check it against your lab scores.

3. How many third-party scripts is too many for an eCommerce site?

The web average is around 20 scripts per page. eCommerce sites frequently exceed this, and each additional script adds main-thread competition. There's no fixed number, but if you're above 25-30 scripts on key pages like product detail or checkout, you likely have redundant or unused integrations worth auditing. The checkout page is the highest-priority target since it's where performance matters most and where script conflicts most directly cost you revenue.