What is a good Core Web Vitals score?

A good Core Web Vitals score means all 3 metrics land in the "good" band - Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, and Cumulative Layout Shift under 0.1 - measured across 75% of your real visits. Get any one of the three wrong and the page fails, no matter how the other two look. The number you see on a speed test out of 100 is a different thing entirely, and confusing the two is the most common reason small-business owners think they have passed when they haven't.

The 3 metrics and their thresholds

Core Web Vitals is exactly 3 measurements, each with three bands - good, needs improvement, and poor. The thresholds are not opinions; Google publishes them and applies them the same way to every site on the web.

Largest Contentful Paint (LCP) measures loading - how long until the biggest thing on screen (usually your hero image or headline) actually appears. Good is 2.5 seconds or less; needs improvement is 2.5 to 4 seconds; poor is anything over 4 seconds. This is the metric most small sites fail, and it is almost always a heavy hero image or a slow host. See what Largest Contentful Paint is for the plain-English version.

Interaction to Next Paint (INP) measures responsiveness - when a visitor taps a button or a menu, how long before the page visibly reacts. Good is 200 milliseconds or less; needs improvement is 200 to 500 milliseconds; poor is over 500 milliseconds. INP replaced the older First Input Delay metric in March 2024 and is harder to fake, because it watches every interaction across the visit, not just the first one. More on what Interaction to Next Paint is.

Cumulative Layout Shift (CLS) measures visual stability - how much the page jumps around as it loads. You have felt this: you go to tap a link and an ad or image loads above it, shoving the link down, and you tap the wrong thing. Good is 0.1 or less; needs improvement is 0.1 to 0.25; poor is over 0.25. CLS is a unitless score, not a time. Here is what Cumulative Layout Shift is in full.

A page passes Core Web Vitals only when all 3 sit in the good band at the same time. There is no average, no "two out of three." A site with a brilliant LCP and INP but a CLS of 0.3 fails outright.

What "75th percentile" means

Here is the part most explanations skip, and it changes how you should read your own numbers. Google does not grade your site on one test run. It grades it on the 75th percentile of your real visits over the trailing 28 days - meaning 75% of the people who actually loaded your page have to hit the threshold for you to pass.

Put plainly: the slowest quarter of your visitors still counts. You don't get to pass on the strength of your own fast laptop on office fibre. If your hero loads in 1.8 seconds for you but 9 seconds for a customer on a 3-year-old Android over patchy regional mobile, that customer is part of the 75th-percentile calculation. The threshold has to hold up for the visitor having a bad day on a cheap phone, because that visitor is the median Australian small-business customer more often than owners like to admit.

Why the 75th percentile and not the median, or the worst case? It is a deliberate compromise. Grading on the median (the 50th percentile) would let a site pass while half its visitors had a poor experience - too soft. Grading on the slowest 1% would punish every site for the one visitor on a decade-old phone in a tunnel - too harsh. The 75th percentile says: three out of four people should have a good experience, and the fourth is allowed to be the edge case. It is a reasonable bar, and most sites that fail it are not failing because of edge cases - they are failing for everyone.

This is also why the field data Google uses (collected from real Chrome users) can disagree sharply with a one-off test from your desk. The test is one fast run; the score Google cares about is the experience of your whole audience, slow phones included.

Score out of 100 vs passing Core Web Vitals

The "92/100" or "47/100" you see at the top of PageSpeed Insights is a lab score - a single simulated load on Google's own hardware and connection. It is useful for diagnosing what to fix, but it is not Core Web Vitals, and a good score does not mean you pass.

They are two separate things:

• The score out of 100 is Lighthouse running one synthetic test, in a lab, right now. It weights several timings into a single number and changes every time you run it.
• Passing Core Web Vitals is the pass/fail verdict from your real visitors' field data over 28 days, judged at the 75th percentile against the 3 thresholds above.

You can score 90+ in the lab and still fail Core Web Vitals in the field, because your real audience is slower than Google's test machine. You can also score in the 70s yet pass, if your real-world experience holds up. When the two disagree, the field verdict is the one Google uses for ranking. Treat the 100-point score as your to-do list and the pass/fail as your report card.

What's realistically achievable for a small site

For a normal small-business site - a few pages, some images, a contact form, maybe a booking widget - passing all 3 is realistic and does not need a developer on retainer. Most failures come from a short, boring list: an oversized hero image, a slow or oversold shared host, a pile of third-party scripts (chat widgets, multiple analytics tags, social embeds), and ad or image slots with no reserved space causing layout shift.

The gap between "realistic" and "where most sites actually are" is wide. A large share of small-business sites sit well past the 2.5-second "good" threshold for LCP - slow enough that a visitor feels it. That is not a hard technical problem; it is a neglected one. Compress the hero, move to a host that isn't oversubscribed, cut the scripts you don't use, and set explicit width and height on images - and most sites move into the good band without a rebuild. Our plain-English Core Web Vitals walkthrough goes through the fixes in order.

If you want to know where your own site actually sits rather than guessing, scan it - the Speed category reports your LCP and the rest in plain language.

When this isn't enough

Passing Core Web Vitals is a floor, not a finish line. Two honest limits worth naming.

First, Core Web Vitals is one ranking input among many, and a fairly minor one. A page that loads beautifully but answers the wrong question still won't rank - speed is a tie-breaker between comparable pages, not a substitute for being the right page. We cover whether Core Web Vitals affect your ranking in detail, and the honest answer is "yes, but less than you'd hope." The fuller picture of how page speed affects Google ranking is worth reading before you spend a fortnight shaving milliseconds.

Second, these 3 metrics measure perceived load experience - and nothing else. They say nothing about whether your site is secure, whether your email can be spoofed, whether your domain is locked, or whether your DNS is configured to keep you online. A site can pass Core Web Vitals with a perfect green score and still leak customer data or hand a scammer the keys to your inbox. Speed is the visible half of being well-run online. The invisible half - the half customers can't see until it fails - is where most of the real risk lives, and it is worth checking with the same rigour you'd give a slow hero image.