Rank Higher Grow Faster

PageSpeed Insights shows “no data” when your page has not had enough visits from Chrome users in the last 28 days for Google to publish real-visitor results. Google only shows field data — data from real people — when there is enough of it to be reliable. A small or new website often has page-level data missing, origin-level data only (an average for your whole domain), or no field data at all. Your lab score — the simulated test run on the spot — is still valid and is what you should work from until field data appears.

That one paragraph resolves most of the panic. “No data” is a traffic statement, not a speed verdict. Google is saying “we don’t have enough real visits to judge,” not “your site failed.”

Three things to hold on to before you read further:

  • The pass/fail that matters for Core Web Vitals comes from field data, not from the 0–100 lab score. When there is no field data, there is no pass/fail yet — in either direction.
  • The lab test is still useful. It is a controlled simulation that finds real problems — oversized hero images, render-blocking scripts, layout shift — before real visitors report them.
  • Fixes do not show up in field data tomorrow. Field data is a rolling 28-day window, so a fix takes up to about 28 days to fully replace the old data (see §9).

✓ Reassurance, with evidence

You are not alone in this. In the HTTP Archive Web Almanac 2025 (July 2025 CrUX data), only 48% of mobile sites passed all three Core Web Vitals at all (56% on desktop). LCP was the most-failed metric — only 62% of mobile sites were “good” on it. Struggling with these numbers puts a small site in the majority, not in a penalty box. Source: HTTP Archive, Web Almanac 2025, Performance chapter — corroborated via citing articles in the research for this cluster; the chapter itself was not opened directly, so treat the figures as corroborated, not primary-verified.

2. The two different reports hiding inside PageSpeed Insights

PageSpeed Insights (pagespeed.web.dev) looks like one report. It is actually two completely different kinds of measurement stacked on one page — and almost every misunderstanding comes from treating them as the same thing.

Field data — “Discover what your real users are experiencing.” Lab data — “Diagnose performance issues”

Field data = real visitors, last 28 days

FIELD

DATA

This top section comes from the Chrome User Experience Report (CrUX): anonymised measurements from real Chrome users who actually visited your pages. It shows LCP, INP and CLS at the 75th percentile — meaning 75% of visits were at or better than the number shown. Google uses this data for the Core Web Vitals assessment. If you don’t have enough traffic, this is the section that says there is no data.

Where it comes fromLighthouse simulation, run when you click “Analyze”CrUX — real Chrome visits over the previous 28 days
The headline numberA single 0–100 Performance scoreLCP / INP / CLS values at the 75th percentile — no single score
Device & networkOne simulated phone, one throttled connection, same every timeYour actual visitors’ phones, networks and locations, all mixed
Changes when…Immediately, every run — and it varies run to run (§6)Slowly — the 28-day window has to roll forward (§9)
Used for Google ranking?No. Diagnostic onlyYes — this is the Core Web Vitals assessment data
Missing when traffic is low?Never — it is generated on demandYes — this is the “no data” you are seeing

Secondary keywords covered in this section: lab data vs field data; PageSpeed Insights explained.

3. Where field data comes from: CrUX, in plain English

CrUX — the Chrome User Experience Report — is Google’s public dataset of how real Chrome users experience websites. Three details explain everything about the “no data” message:

It only counts Chrome visits, with conditions

The visit has to happen in Chrome, the user has to have usage-statistics sharing on, and the page has to be publicly reachable and indexable. Visits in Safari, in apps, or from your own logged-in WordPress admin sessions do not build your CrUX data. A store whose customers mostly arrive via the Instagram app’s in-app browser or Safari on iPhones can have real sales and still have thin CrUX data.

It is judged at the 75th percentile, over 28 days

Google does not use your average, and it does not use your best day. It takes a rolling 28-day window and looks at the point where 75% of visits were that fast or faster. One bad week — a broken plugin, a traffic spike from a slow network — stays in your field data for weeks after you fix it. That is frustrating, but it is also fair: it stops one lucky test from declaring a slow site “fast.”

Page data first, origin data as the fallback

PageSpeed Insights tries to show data for the exact URL you tested. If that URL does not have enough visits, it falls back to origin-level data — the aggregate for your entire domain — and labels it as such. Read that label carefully: origin data mixes your fast homepage with your slowest product page, so it can hide a problem on one template or invent one that a single page doesn’t have. If even the origin doesn’t have enough visits, you get no field data at all.

Tool change — Nov 2025

The old CrUX Dashboard (the Looker Studio report many older tutorials link to) was retired at the end of November 2025. If a guide sends you there, it is out of date. Use CrUX Vis or the CrUX History API for trend data instead. Source: Chrome for Developers, “CrUX Dashboard deprecation,” 9 Sep 2025.

4. The four reasons a small website gets “No Data”

  • Not enough Chrome visits — the usual cause. Google does not publish the exact minimum visit count, and you should be suspicious of any article that quotes one as fact. What is confirmed is the principle: below a sufficiency threshold, the data is withheld rather than shown unreliably. New sites, local businesses, and niche stores sit below it for a long time — often on every page except the homepage.
  • The page is new or the URL changed. A page needs its own visit history. Publish a new product, change a slug, or move from http to https (or add/remove www) and that URL starts again from zero — even if the old version had data.
  • Traffic is spread too thin. A store with 400 products may get healthy total traffic while each product page gets a handful of Chrome visits a month. The homepage gets field data; the pages that actually sell stay blank. This is why origin-level fallback data exists — and why you should test by template (one product page, one category page, the checkout) in the lab, instead of waiting for every URL to earn field data.
  • Your visitors aren’t counted. Heavy iPhone/Safari audiences, traffic from in-app browsers, or visitors who decline telemetry all shrink the counted pool below your real visitor numbers. “No data” can coexist with a busy shop.

Notice what is not on this list: a Google penalty, an indexing block, or a broken site. A page can be indexed, ranking, and selling — and still show no field data. (If your page is also not indexed, that is a separate problem — check it in Search Console’s Page Indexing report, not in PageSpeed Insights.)

Secondary keyword covered here: Search Console Core Web Vitals “not enough data” — Search Console’s Core Web Vitals report draws on the same CrUX data, so low-traffic sites see “no data” or very few URLs there for the same reason, and for the same non-alarming cause.

5. Which score actually matters — and for what

Here is the hierarchy, stated as plainly as Google states it:

What does Google use for the Core Web Vitals ranking assessment?Field data (CrUX) — LCP, INP and CLS at the 75th percentile, assessed separately for mobile and desktop.
What is the 0–100 score for?Diagnosis. It summarises one lab run so you can see whether a change helped, and it comes with the “Opportunities” and “Diagnostics” lists that tell you what to fix.
Does a bad lab score directly lower my rankings?No. But it often predicts bad field data later — the same heavy image that fails the lab test will slow real visitors too. Treat it as an early warning, not a sentence.
How big a ranking factor is Core Web Vitals, honestly?Real, but modest. Google “highly recommends” good Core Web Vitals and says they align with what its core ranking systems reward — and its page-experience guidance is equally clear that the most relevant content can still rank well with sub-par page experience. Speed will not rescue irrelevant content, and slow speed alone rarely buries strong content. Fix it because visitors leave slow pages (§7), not out of ranking panic.
Does page speed decide whether AI search (AI Overviews, ChatGPT, Perplexity) cites me?Google states there are no additional requirements to appear in AI Overviews or AI Mode beyond being indexed and eligible to show with a snippet — and page speed is not a stated direct criterion for AI citation. For a small site, getting indexed is the AI-visibility problem; speed helps only indirectly, through ordinary ranking. Any article claiming speed is an “eligibility gate” for AI citations is going beyond Google’s documentation.

Sources: Google Search Central, “Understanding Core Web Vitals and Google search results” (updated 10 Dec 2025, per search result); Google Search Central, “AI features and your website.”

6. Why your score changes every time you run it

You run the test at 9 am: 71. Again at 9:05: 58. After lunch: 77. Nothing on the site changed. This is the second question every owner asks — and it also has a boring, reassuring explanation.

The lab test loads your live page, once, over a simulated network. Real conditions leak in around the simulation: your server answers a little slower under load, a third-party script (chat widget, analytics, a review badge) responds late or not at all, an ad or embed rotates in something heavier, a shared-hosting neighbour spikes at the same moment. LCP is especially jumpy because it depends on whichever large element wins the race to paint on that particular load.

How to read a moving score

  • Run the test three times and take the middle result. One run is an anecdote.
  • Compare the metrics, not just the score. The 0–100 score is a weighted blend; LCP in seconds tells you more than a 6-point swing.
  • Test mobile first. The mobile simulation is deliberately harsh — a mid-range phone with roughly a 4× CPU slowdown on a ~1.6 Mbps simulated slow-4G connection, versus desktop’s faster setup — which is why mobile scores sit far below desktop on the same page. That gap is the subject of the companion article in this cluster, “Why Is My Website Slow on Mobile but Fast on Desktop?”
  • Test at a quiet time before blaming a change you made. If all three runs drop right after a change, that’s a signal. If they wander, that’s noise.

Secondary keyword covered here: why does my PageSpeed score change every time I run it.

7. Do you need a 100/100 score? No — you need these three numbers

No small business needs 100/100, and chasing it wastes money: the last few points usually come from removing things that make you money (analytics, chat, reviews, payment badges). What Google actually assesses is whether your field data clears three thresholds at the 75th percentile. These are unchanged for 2025–2026:

LCP — Largest Contentful PaintHow fast the main content (usually your hero or product image) appears≤ 2.5s2.5–4.0s> 4.0s
INP — Interaction to Next PaintHow fast the page responds when someone taps or clicks≤ 200ms200–500ms> 500ms
CLS — Cumulative Layout ShiftHow much the page jumps around while loading≤ 0.10.1–0.25> 0.25

How the Core Web Vitals metrics thresholds were defined” (published 21 May 2020, last updated 7 May 2025 — opened directly for this cluster’s research, 3 Nov 2026). A related number worth knowing: TTFB (time to first byte) is not a Core Web Vital, but web.dev advises striving for ≤0.8s (poor is >1.8s) because a slow server delays everything after it.

And why chase “good” at all, if not for the score? Because visitors behave measurably better on faster pages. In Deloitte Digital’s Milliseconds Make Millions study for Google (2020; 37 brands, ~30 million sessions), a 0.1-second improvement in mobile speed was associated with +8.4% retail conversions and +9.2% average order value — with the honest caveat, which the study itself carries, that the gains came when speed improved across the whole journey, not just the homepage. In Portent’s research (2022; 100M+ page views), an ecommerce page loading in 1 second converted at 3.05% versus 1.68% at 2 seconds. You do not need a perfect score to capture most of that difference — you need to not be slow.

Secondary keyword covered here: do I need a 100 PageSpeed score, small business.

8. What to do when you have no field data: a 6-step working method

  • Confirm it’s a data problem, not a site problem. Open the page in an incognito window on your phone. If it loads and works, your site is fine — it’s the dataset that’s empty. Then check Search Console → Indexing → Pages to confirm the URL is indexed.
  • Work from lab data — by template, not by homepage. Run PageSpeed Insights (mobile) on five URLs: homepage, a best-selling product page, a category/collection page, the blog post that gets the most traffic, and the checkout (lab-test it with a product in the cart in your own browser tools if PSI can’t reach it). Templates share the same images and scripts, so fixing one product page’s LCP usually fixes them all.
  • Fix in this order: LCP element, then layout shift, then the heaviest scripts. In the lab report, find the “LCP element” named in the diagnostics — on a store it is almost always the main product or hero image. Make sure it is right-sized and compressed, that it is not lazy-loaded, and that it carries fetchpriority=”high”. Then fix anything that makes the page jump (unsized images, banners that push content down). Then look at the largest third-party scripts.
  • Use Search Console’s Core Web Vitals report for what it can show. With low traffic, it may group most URLs as having insufficient data — that’s expected. What matters is that it groups by template, so when data does arrive, issues arrive already organised (“LCP issue: longer than 2.5s on 14 URLs”).
  • Watch trends in CrUX Vis, not the retired dashboard. CrUX Vis shows your origin’s field-data history when enough exists. Check monthly, not daily — the 28-day window makes daily checking meaningless.
  • Re-test the lab score after each fix; re-check field data after 28 days. Lab = “did my change work?” (immediate). Field = “did Google’s assessment change?” (up to 28 days — next section). Keeping those two questions separate stops 90% of PageSpeed confusion.

✓ The mindset shift

No field data means you are pre-assessment, not failing. The work you do now, guided by lab data, is what decides whether your first real field-data verdict — when traffic grows enough to produce one — is a pass.

9. The 28-day rule: why fixes “don’t work” (they do — slowly)

Because field data is a rolling 28-day window, today’s field numbers still contain up to 27 days of visits to the old, slow version of your page. Even a perfect fix only replaces roughly 1/28th of the window per day. In practice:

  • Days 1–7: field data barely moves. Your lab score is the only honest feedback.
  • Days 7–21: the 75th-percentile numbers start to slide toward the new reality — if real visitors’ experience actually improved.
  • Day 28: the window is fully post-fix. Only now is a pass/fail judgement on the new version complete. Search Console additionally wants a sustained pass across its own validation period before it marks an issue “fixed.”

Two practical consequences: never judge a speed fix by next morning’s PageSpeed field section, and never pay someone who promises your Core Web Vitals will “pass this week.” The calendar does not negotiate.

10. What actually changed in 2025–2026 (and what didn’t)

Thresholds: LCP ≤2.5s, INP ≤200ms, CLS ≤0.1UnchangedAny article announcing “new 2025 limits” is wrong. The targets in §7 are current.
INP replaced FID as a Core Web Vital12 Mar 2024FID is a dead metric. If your theme, plugin or a freelancer still reports FID, the advice is out of date.
LCP & INP measurable in all major browsers (Safari 26.2 completed the set; Firefox added INP in the same Interop 2025 cycle)Dec 2025Your own analytics tools can now measure iPhone users too. But CrUX — the data Google assesses — remains Chrome-only, so your PageSpeed/Search Console numbers do not change because of this.
CrUX Dashboard retired; use CrUX Vis / History APIEnd of Nov 2025Update any bookmarked tutorial links (§3).
Google on AI search eligibilityCurrent guidance, checked Nov 2026No extra optimisation and no speed requirement for AI Overviews/AI Mode: be indexed, be snippet-eligible (§5).

Could not verify — stated plainly

Two things circulating in 2026 blog posts were not confirmed against Google’s own pages in the research behind this article, so this article does not rely on them: a Search Console “Generative AI performance report” (reported by secondary sites as rolling out 31 Aug 2026; Google’s AI-features page still describes AI traffic inside the “Web” search type) and a PageSpeed Insights “Agentic Browsing” category (single-blog source). If either matters to you, check Search Console / PageSpeed Insights directly rather than trusting a blog summary — including this one.

11. Seven mistakes small businesses make with PageSpeed Insights

  • Judging the site by the desktop score. Google assesses mobile and desktop separately; most visitors are mobile, and the mobile test is the harsher one. Mobile is the number to work on.
  • Treating one lab run as truth. Three runs, take the middle (§6).
  • Panicking at “no data.” It is a traffic threshold, not a failure (§1, §4).
  • Only ever testing the homepage. Product and checkout templates carry the revenue — test those (§8).
  • Deleting money-making features to chase 100. Thresholds, not perfection (§7).
  • Checking field data daily after a fix. The window is 28 days (§9). Check lab now, field monthly.
  • Reading origin data as page data. Check the label — origin data averages your whole domain (§3).

12. Frequently asked questions

These are the exact questions owners type into Google, ChatGPT and Perplexity on this topic — answered in the same words, so this page can be the cited answer.

Why does PageSpeed Insights say “no data” for my site?

Because your page (or your whole domain, at origin level) has not received enough visits from Chrome users in the last 28 days for Google to publish reliable real-visitor data. It is a traffic threshold, not a speed failure. Use the lab results further down the same report to guide fixes until field data appears.

Next in this cluster: “Why Is My Website Slow on Mobile but Fast on Desktop?” (the diagnosis hub) and “Deleted Shopify Apps Can Still Slow Your Store.” Back to the pillar: the complete guide to website page speed and its effect on sales and revenue on Techish Web.

Leave a Reply

Your email address will not be published. Required fields are marked *

Categories
Our Services
Contact Us
About Us

TechishWeb is your trusted source for technology, SEO, digital marketing, web development, AI insights, and business solutions. We help businesses grow online with innovative strategies and expert digital services.