In short
A good page shows its main content within 2.5 seconds, responds to a tap or click within 200 milliseconds, and does not shift its layout while loading; those are Google's Core Web Vitals thresholds, measured for real visitors at the 75th percentile. In practice, a page whose main content appears in under 1.5 seconds on a mid-range phone feels fast, 2.5 seconds feels acceptable, and beyond 4 seconds a meaningful share of visitors leave. The old single 'fully loaded' time matters less than when the page becomes usable.
Why 'load time' is the wrong question
A page is not loaded at one moment. The first byte arrives, then the layout, then the main image, then the fonts, then the scripts that make buttons work, then the analytics and the chat widget, and finally the last lazy-loaded image at the bottom of the page that nobody has scrolled to. "Fully loaded" is the last of those, and it is the one the visitor cares least about, because by then they have been reading for several seconds.
What visitors experience is a sequence of questions: is something happening? Can I see what I came for? Can I use it? Did it stop moving? The modern metrics each answer one of those.
The numbers that replaced it
Time to First Byte (TTFB). How long until the server starts responding. Under 800 milliseconds is Google's threshold; under 200 is what a well-configured site achieves. Everything else waits on this, so it is the first thing to check when a site is slow everywhere.
Largest Contentful Paint (LCP). When the largest visible element, usually the hero image or the main heading, has rendered. This is the closest thing to what people mean by "load time". Google's thresholds: good under 2.5 seconds, needs improvement to 4 seconds, poor above.
Interaction to Next Paint (INP). How long the page takes to respond visibly when a visitor taps, clicks or types, measured across the whole visit. Good under 200 milliseconds, poor above 500. This replaced First Input Delay in 2024 and it is where JavaScript-heavy sites fail.
Cumulative Layout Shift (CLS). How much the layout jumps while loading. Good under 0.1. Not a time, but the thing that makes a page feel broken even when it is fast.
These three, LCP, INP and CLS, are the Core Web Vitals. Google measures them from real Chrome visitors over 28 days and uses the 75th percentile, meaning three-quarters of visits must meet the threshold for the page to pass.
What the research says about patience
The often-quoted figures are that a one-second delay costs a few percent of conversions, that bounce probability rises sharply as load time passes three seconds, and that most mobile visitors abandon a page that takes longer than about three seconds to show content. The exact numbers vary by study and by decade, and some of the most-cited ones are old, but the shape is consistent and repeatedly confirmed: the relationship between speed and abandonment is steep between one and four seconds, and flattens after that because the impatient have already gone.
There is also a perception threshold around 100 milliseconds below which a response feels instant, and around one second below which the visitor's flow of thought is not interrupted. Those are the numbers behind the INP threshold and the reason "fast enough" is stricter for interactions than for initial load.
Sensible targets by site type
Marketing and content sites. LCP under 1.5 seconds on a mid-range phone over 4G, because these visitors have the least commitment and the most alternatives. This is achievable with a static build, optimised images and few scripts, and is the target we set for our own work.
E-commerce. LCP under 2 seconds and INP well under 200 milliseconds, because every product page is a landing page and every filter interaction is a chance to lose the visitor. Third-party scripts are the usual obstacle.
Web applications behind a login. The initial load can be slower, since the user has committed, but INP matters more than anywhere else: an application that lags on every click is unusable regardless of how fast it first appeared.
Anything on a slow connection or old devices. If your visitors are in markets with cheaper phones and slower networks, halve the targets. The Core Web Vitals thresholds are worldwide averages and are generous for those conditions.
How to measure it honestly
Lab tools such as Lighthouse and WebPageTest give repeatable numbers under simulated conditions and explain what to fix. Field data, from Search Console's Core Web Vitals report or the Chrome UX Report section of PageSpeed Insights, tells you what real visitors experienced and is what Google ranks on. The two disagree often, and when they do, the field data is the truth. Real-user measurement on your own site closes the gap.
Measure on a phone, on a mobile connection, from the region your visitors are in, with the browser cache cleared. A developer's laptop on office wifi is the least representative device in the building.
The one-line answer
Main content visible in under 2.5 seconds for three-quarters of real visitors, with under 1.5 as the target for a site that competes on first impressions; interactions under 200 milliseconds; no layout shift. Meet those and speed has stopped being the reason anyone leaves. If you are not sure where your site sits, book a call and we will pull the field data together.
Common questions
What is a good website load time in 2026?
Main content visible within 2.5 seconds for at least three-quarters of real visitors is Google's threshold; under 1.5 seconds is a strong target for marketing and content sites. Interactions should respond within 200 milliseconds and the layout should not shift while loading. The old 'fully loaded' time matters far less than when the page becomes usable.
Is a 3-second load time good?
It is in Google's 'needs improvement' band for Largest Contentful Paint, which runs from 2.5 to 4 seconds. Visitors tolerate it, but abandonment rises steeply between one and four seconds, so a page at three seconds is losing a measurable share of people that a page at 1.5 seconds keeps. It is usually one or two fixes from good.
What is the difference between load time and Largest Contentful Paint?
Load time traditionally meant when every resource on the page had finished downloading, including scripts and images nobody has scrolled to. Largest Contentful Paint measures when the largest visible element has rendered, which is when the visitor can see what they came for. LCP is the better proxy for experience and is what Google measures.
How fast should a mobile page load?
The same thresholds apply, but they are harder to meet: LCP under 2.5 seconds on a mid-range phone over a mobile connection, which is what Google's field data mostly consists of. Oversized images and JavaScript are the usual reasons a page that is fast on desktop misses the mark on phones.
Does page load time affect SEO?
Yes. Core Web Vitals, measured from real visitors, are a ranking signal, and Google indexes the mobile version first. The ranking effect is modest compared with relevance and links, but the effect on visitors leaving before the page is usable is not, and that shows up in every other metric.
