Storefront Speed Journal

Reviewed November 14, 2026

How to Read a PageSpeed Insights Report for Your Store

A PageSpeed Insights report has two sections, and they answer different questions. The top section is real-visitor data, Core Web Vitals measured on actual shoppers over roughly four weeks, and that is the verdict: your LCP, INP, and CLS against the thresholds of 2.5 seconds, 200 milliseconds, and 0.1. The bottom section is a lab score from a simulated load, and that is the diagnosis: the opportunities list tells you what to fix. Read the verdict first, then use the diagnosis to fix it. A lab score that changes from run to run is normal. Real-visitor data is what counts.

A store owner runs PageSpeed Insights, sees a mobile score of 42, and panics. Or sees a score of 91 and relaxes. Both reactions miss the point, because the score is not the verdict. The report has two sections, and they answer different questions, and most people read the wrong one first.

The top of the report is real-visitor data: what actual shoppers on actual phones experienced over roughly the last four weeks. The bottom is a lab test: a simulated load on a simulated phone, run once, right now. The lab score changes from run to run. The real-visitor data does not, and it is the number Google uses. Read the top first.

The top: real-visitor data

The top section shows Core Web Vitals from real visitors, when your store has enough traffic for Google to report them. Three numbers: Largest Contentful Paint, how long the main content takes to appear, good at 2.5 seconds or less. Interaction to Next Paint, how quickly the page responds when someone taps a variant or adds to cart, good at 200 milliseconds or less. Cumulative Layout Shift, how much the page jumps while loading, good at 0.1 or less. Google wants at least 75 percent of visits to meet each target, and it measures mobile and desktop separately, because the two often tell different stories.

If the top section says "not enough data," your store does not have the traffic for real-visitor metrics yet, and the lab score is all you have. That is fine, but treat it as a rough guide, not a verdict. And remember that real-visitor data covers a rolling period of roughly four weeks, so improvements take time to appear. Fix something today, and the real-visitor number moves over the next month, not tomorrow.

The bottom: the lab score and the opportunities

The lab score, 0 to 100, is a weighted summary of a simulated load. It is useful as a before-and-after comparison on the same page, and misleading as an absolute judgment. A score of 42 on a page with a big hero image and a score of 91 on a text page are not comparable, because the pages are doing different jobs. What matters in the lab section is the opportunities list, which names the specific problems in order of impact: oversized images, render-blocking scripts, unused JavaScript, a slow server response.

Work the opportunities in order, biggest first, and re-run the report after each fix. The list is the diagnosis, and it is specific: it names the file, the size, and the saving. A lab score that moves 10 points after a fix is a signal, not a result. The result is the real-visitor LCP moving under 2.5 seconds over the next four weeks.

The three pages to test, and the trap

Test three pages: the homepage, a busy collection page, and your best-selling product page, all on mobile. The homepage is usually the slowest, because it carries the hero image, the sliders, and the most apps. The product page is the one that matters most for revenue, because it is where the purchase happens. A store that only tests its homepage is measuring its slowest page and ignoring the page that earns money.

The trap is chasing the lab score instead of the real-visitor data. A theme change that raises the lab score by 20 points but does not move real-visitor LCP has changed a simulation, not a store. And a store with a lab score of 60 and real-visitor LCP of 2.1 seconds is faster, for its actual customers, than a store with a lab score of 90 and real-visitor LCP of 3.4 seconds. The score is a tool. The visitors are the verdict.

Related reading

Two earlier posts on this site cover ground “How to Read a PageSpeed Insights Report for Your Store” builds on: Your Product Images Are 56 Times Too Big: Sizing Rules That Move LCP, Video Hero LCP: The Poster-Image Facade That Cut 9 Seconds to 2.7, and Lazy Loading Below the Fold: Where It Helps LCP and Where It Hurts.

Frequently asked questions

What is the difference between the lab score and real-visitor data?

Real-visitor data is Core Web Vitals measured on actual shoppers over roughly four weeks, and it is the verdict Google uses. The lab score is a simulated load run once, right now. The lab score changes from run to run. Real-visitor data is what counts.

What are good Core Web Vitals for a Shopify store?

LCP of 2.5 seconds or less, INP of 200 milliseconds or less, and CLS of 0.1 or less, met by at least 75 percent of visits. Check mobile and desktop separately, since the two often tell different stories.

Which pages should I test in PageSpeed Insights?

Three: the homepage, a busy collection page, and your best-selling product page, all on mobile. The homepage is usually slowest, and the product page is where the purchase happens.

Why does my PageSpeed score change between runs?

Because the lab test is a simulation, and network conditions, server load, and the test environment vary each time. A lab score that moves a few points between runs is normal. Judge fixes by the real-visitor data over four weeks, not by a single lab run.