Core Web Vitals 101: what actually affects load time
Core Web Vitals are Google’s three measures of how a page is experienced: LCP for when the main content appears, INP for how quickly the page responds to input, and CLS for how much the layout shifts. They are measured on real visitors rather than in a test tool, and it is the 75th percentile that decides whether you clear the threshold. A score of 95 in a lab test therefore says almost nothing about what your visitors experience.
What are the thresholds?
Three metrics, three thresholds, and one most common cause each. The limit applies to the 75th percentile of real visits, not to a single test.
| Metric | Good | Needs work | Most common cause |
|---|---|---|---|
| LCP | under 2.5 s | over 4.0 s | Unoptimised hero image or render-blocking font |
| INP | under 200 ms | over 500 ms | Third-party scripts blocking the main thread |
| CLS | under 0.1 | over 0.25 | Images and ads without reserved space |
What is LCP and what fails it?
Largest Contentful Paint measures when the largest element above the fold has been painted, usually an image or a heading. The threshold is 2.5 seconds. In practice it is almost always one of three things:
- An image served at full resolution while displayed at a quarter of the size.
- A font blocking render, so the text exists but is not visible.
- Content assembled in the browser, so nothing exists until JavaScript has run.
What is INP and what fails it?
Interaction to Next Paint replaced the old FID metric and is stricter. It measures how long from a click until something visibly happens, across the whole visit. The threshold is 200 milliseconds. The cause is almost always JavaScript blocking the main thread: analytics scripts, chat widgets, A/B testing tools. Each one is a cost, and they add up.
What is CLS and what fails it?
Cumulative Layout Shift measures how much content moves after being painted. The threshold is 0.1. This is the easiest metric to fix and the most irritating to experience: an ad loading in and pushing the body copy down just as you started reading. The fix is always the same: reserve the space in advance with width and height on images, and an explicit height on anything loaded later.
Why does a caching plugin not help?
A caching plugin improves server response time, the delay before the page starts being delivered. That is rarely what fails Core Web Vitals. LCP is decided by images and fonts, INP by JavaScript, CLS by unreserved space. None of them is solved by a cache in front of the page. Caching is good, but it is not the same thing as performance.
Why is a lab test not enough?
Lab tests run on a fast machine with a perfect connection. Field data comes from real visitors on real phones. The Chrome UX Report gives you field data free if you have enough traffic. If you do not, measure in the visitor’s browser yourself. Sites that look good in the lab and poor in the field almost always have a third-party script problem.
How much does it affect rankings?
Honestly: a little. Core Web Vitals are part of Google’s page experience signals but weigh lightly next to content and relevance. The real gain is commercial. Faster pages lose fewer visitors before they have seen anything, and that shows up in bounce rate and conversion far more clearly than in positions.
Common questions
How fast does a page have to load to pass Core Web Vitals?
LCP under 2.5 seconds, INP under 200 milliseconds and CLS under 0.1, measured at the 75th percentile of real visits over 28 days. A single test showing good numbers is therefore not enough: a quarter of visits may be slower, but no more.
Why does PageSpeed Insights show different numbers every time?
The tool shows two things at once. The lab run is repeated on every test and varies with server load and network. The field data is built from 28 days of real visits and moves slowly. Field data decides whether you clear the threshold, and it is the one worth tracking.
We do not have enough traffic for field data. What then?
Measure in the visitor’s browser yourself using the built-in browser API for the three metrics, and send the numbers to your own log. That gives the same kind of data as the Chrome UX Report but for your own traffic at any volume, and it is a few lines of code rather than a tool purchase.
Do Core Web Vitals really affect rankings?
They are part of Google’s page experience signals but weigh lightly next to content and relevance. The bigger gain is commercial: faster pages lose fewer visitors before they have seen anything, and that shows up in bounce rate and conversion far more clearly than in positions.
More in the same area
Replatforming without losing traffic: a checklist
What actually decides whether traffic survives a rebuild, in the order the work should happen.
API integrations: six mistakes that cost money
The errors that make an integration work in testing and then create duplicates, lose orders, or need daily babysitting.
