What INP measures
INP is the Core Web Vital for responsiveness. Each time someone clicks a button, taps a menu or types in a field, the browser has to run your code and draw the result. INP measures the time from the input to the next frame painted on screen, across every such interaction during the visit, and reports the longest one, ignoring outliers.
Only clicks, taps and key presses count. Hovering, zooming and scrolling are not measured. A page with no interactions has no INP.
INP replaced First Input Delay (FID) as a Core Web Vital in March 2024. FID only measured the wait before the first interaction's code started; INP covers the whole response, for every interaction.
Good, needs improvement and poor
| Rating | INP |
|---|---|
| Good | 200 ms or less |
| Needs improvement | Over 200 ms, up to 500 ms |
| Poor | Over 500 ms |
As with the other Core Web Vitals, judge a page at the 75th percentile of real page loads, split by mobile and desktop.
Example
On a pricing page, clicking "Yearly" re-renders every plan card and recalculates prices. On a fast laptop it takes 80 ms. On a mid-range phone it takes 450 ms, and the visitor taps twice because nothing seemed to happen. If a quarter of your page loads come from such phones, the page's p75 INP lands in "needs improvement". Updating only the price text instead of re-rendering the cards brings the phone case down to around 120 ms.
How to measure INP
INP needs real interactions, so field data matters most. Collect it from real visitors with the browser's performance APIs (Google's web-vitals library is the common choice) or read it from the Chrome UX Report. In the lab, Chrome DevTools can record an interaction and show which long task delayed it. The usual fixes are breaking up long JavaScript tasks, doing less work in event handlers, and avoiding large layout and style recalculations after a click.
Common INP mistakes
- Testing only on a fast machine. INP problems show up on mid-range phones.
- Expecting a lab score. Lighthouse has no INP because it doesn't click; it reports Total Blocking Time as a stand-in.
- Measuring only the first interaction. That was FID. A slow "Add to cart" late in the visit counts too.
- Blaming third-party scripts without checking. Record the interaction and see whose code ran.
How MetricRaven measures INP
With Page speed turned on in Settings (it is off by default, on every plan), MetricRaven's add-on uses Google's web-vitals library and sends one snapshot per page load with INP, LCP, CLS, FCP and TTFB, as far as the browser measured them.
Pages → view all shows an INP column at the 75th percentile with a green, amber or red dot, next to purchase rate. It appears from 1024 px wide (1280 px while Scroll depth is on); on narrower screens one dot shows the worse of the LCP and INP ratings. INP is also half of the slow-versus-fast comparison: a visitor with any load of a page over 500 ms INP (or over 4 s LCP) counts as slow, and the API lists pages where slow visitors paid less often than fast ones. That is correlation, not proof. See Page speed vs conversion.
Questions
- What is a good INP score?
- 200 milliseconds or less at the 75th percentile of page loads. Between 200 and 500 milliseconds needs improvement, and over 500 milliseconds is poor.
- When did INP replace FID?
- INP became a stable Core Web Vital in March 2024, replacing First Input Delay, which only measured the delay before the first interaction was handled.
- Does scrolling affect INP?
- No. INP observes clicks, taps and key presses. Hovering, zooming and scrolling are not measured, although a tap that starts a scroll gesture can be.
- Why does my page have no INP value?
- INP needs at least one click, tap or key press. Visitors who only read and scroll produce no INP, and some browsers don't measure it at all.