ShipworkSite inspection
Checks
Every error, explainedGuides →
Pricing Learn Guides

What do real visitors experience on your site?

Lab scores measure one synthetic run. The Chrome UX Report measures what actual Chrome users experienced over the last 28 days, and that is the data behind the page-experience signal.

Runs the Real-user Core Web Vitals check against a bounded sample of your site, from outside your network. Nothing is stored.

What this check inspects

Shipwork queries the Chrome UX Report for your origin and reads the 75th percentile of Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift. It reports each metric with a verdict, the share of mobile traffic, and the collection period. Origins without enough Chrome traffic have no data, and it says so rather than inventing a number.

What a failure means

A metric rated poor is an error and one rated needs improvement is a warning. Because the verdict is at the 75th percentile, it reflects the slower quarter of your visitors, not the average. Mobile share matters: if most traffic is on phones, mobile performance decides the verdict.

How to fix it

  1. Fix the largest element first. LCP is usually a hero or product image, so serve it in a modern format at the right size and do not lazy load it.
  2. Reserve space for images and embeds so Cumulative Layout Shift drops.
  3. Reduce main-thread work from third-party scripts to improve Interaction to Next Paint.
  4. Re-measure after each change. Field data updates on a rolling window, so it takes days to reflect a fix.

A typical failure, worked through

The setupA store serves a 1.8MB hero image as a PNG at desktop dimensions on mobile, and injects a chat widget that loads before the image.

What the check reportsLCP at the 75th percentile is poor on mobile, while desktop is good. The check reports the mobile share that makes this the metric that matters.

The pointField data is not a lab score, so a change that looks good in a single test may not move the p75. Target the slowest quarter of real visits.

Questions

Why does it say no data?
The Chrome UX Report only includes origins with enough real Chrome traffic. A new or low-traffic site has none yet, which is normal.
Is the 75th percentile the average?
No. It is the value that 75 percent of visits beat, so it deliberately reflects the slower quarter. Google uses p75 for the page-experience verdict.
How long after a fix does it update?
The report uses a 28-day rolling window, so improvements appear gradually over days and weeks rather than immediately.

Related checks and guides