INP Is the Core Web Vital Nobody Prepared For (Here’s How to Fix It)

When Google swapped FID for INP in March 2024, sites that passed Core Web Vitals for years suddenly started failing. Here’s why, and what you can actually do about it.

SP

SAT Team

6 min read

On March 12, 2024, Google quietly pulled off one of the biggest shake-ups in Core Web Vitals history. First Input Delay (FID) was retired, and Interaction to Next Paint (INP) took its place. Sites that had passed Core Web Vitals for years woke up failing. If that happened to you, you're in good company. Industry crawls at the time suggested that a large share of sites passing FID did not pass INP.

The frustrating part? Most site owners had no idea why. Their pages looked the same. Nothing had changed in the code. And yet Search Console started flagging URLs left and right.

What INP actually measures (and why FID was too easy)

FID only measured one thing: the delay before the browser could start processing your very first interaction. That's it. It ignored everything after the first click, and it ignored how long the actual work took. Almost everyone passed. It was a bit like a driving test that only checked whether you could open the car door.

INP is much tougher. It watches every click, tap, and keypress during the entire page visit, measures how long each one takes from input to the next visual update, and then reports roughly the worst one. So if your menu opens instantly but your "Add to Cart" button freezes the page for 800ms, INP catches it. FID never did.

The thresholds are strict:

  • Good: 200 milliseconds or less
  • Needs improvement: between 200ms and 500ms
  • Poor: over 500ms

And remember, this is measured at the 75th percentile of real users. Your fast laptop on office wifi doesn't count for much. The metric cares about a mid-range Android phone on a spotty connection.

The usual suspects behind bad INP

After auditing hundreds of sites, the same culprits show up again and again.

Long JavaScript tasks. The browser's main thread can only do one thing at a time. If a script runs for 400ms without a break, any tap during that window just sits there waiting. Analytics bundles, A/B testing tools, and tag managers are repeat offenders.

Heavy event handlers. A click handler that recalculates the whole cart, updates the DOM in ten places, and fires three tracking pixels before the browser can paint? That's your INP score right there.

Third-party scripts. Chat widgets, session recorders, ad scripts. Every one of them competes for the same main thread as your buttons.

Rendering too much at once. Frameworks re-rendering giant component trees on a single state change is a modern classic. React devs, look at your context providers.

How to actually fix it

Start by finding your slow interactions instead of guessing. Chrome DevTools has a Performance panel with an interactions track. Even easier: install the official Web Vitals extension, turn on console logging, and click around your page like a normal user. It will print every interaction and its duration.

Once you know which interactions hurt, work through these:

1. Break up long tasks. Anything over 50ms blocks interactions. Split big loops and heavy functions with scheduler.yield() or a setTimeout of zero so the browser can squeeze in pending input between chunks.

2. Give visual feedback first, work later. When a user taps a button, change its state immediately, then do the expensive stuff. Users judge responsiveness by paint, not completion. A spinner that appears in 50ms feels faster than a result that appears in 400ms with nothing in between.

3. Debounce what happens on input. Search-as-you-type fields firing an API call and re-render on every keystroke are INP poison. Wait for a pause in typing.

4. Audit your third parties ruthlessly. Load a copy of your page with the tag manager blocked and measure again. The difference is often shocking. Delay non-essential scripts until after the first interaction, or load them in a web worker with a tool like Partytown.

5. Reduce DOM size. Pages with 5,000+ DOM nodes make every style recalculation slower, which drags out every interaction. Prune what you don't need, and virtualize long lists.

Don't obsess over lab scores here

One thing that trips people up: Lighthouse alone can't fully measure INP because a lab run doesn't click anything by default. Your PageSpeed Insights lab section shows Total Blocking Time (TBT) as a proxy. The INP number you should trust lives in the field data at the top of the report, pulled from the Chrome UX Report, and it needs 28 days of real traffic to update.

So when you ship a fix, don't panic that Search Console still shows failing URLs the next morning. The window moves slowly. Track your TBT in the lab as an early signal, watch field INP over the following month, and use the "Validate fix" button in Search Console once your real-user numbers cross under 200ms.

Common questions about INP

Does INP affect rankings directly? Core Web Vitals are part of Google's page experience signals, and INP is one of the three vitals. It's a tiebreaker more than a dominator: brilliant content with poor INP will still often outrank thin content with great INP. But in competitive niches where everyone's content is decent, experience metrics start deciding who gets the click. And the indirect effect is bigger anyway: sluggish interactions increase abandonment, and lost engagement costs more than any ranking wobble.

My INP is fine on desktop but fails on mobile. Why? Because mobile CPUs are slower and your JavaScript takes two to four times longer to execute on a mid-range phone. Same code, weaker hardware. This is normal and it's why you should always debug with CPU throttling turned on in DevTools, four times slowdown at minimum.

Can a cache plugin or CDN fix INP? No, and this surprises people. Caching fixes server response time and helps LCP. INP happens after the page has loaded, inside the visitor's browser, while your JavaScript runs. The fix is always about doing less work on the main thread at interaction time. No amount of edge caching changes what executes on the device.

What's a realistic target? Under 200ms at the 75th percentile to pass. If you're starting from 500ms or worse, aim for staged wins: get under 400ms first by removing the worst third-party scripts, then chase the long tasks in your own bundle. Sites rarely jump from poor to good in one release, and that's fine.

The honest takeaway

INP forced a reckoning that was overdue. For years, sites shipped megabytes of JavaScript and passed Core Web Vitals anyway because FID was so forgiving. That loophole is closed.

The good news is that INP fixes tend to be felt by real people. Faster menus, snappier carts, forms that respond instantly. You're not gaming a metric; you're removing friction that was quietly costing you conversions all along. Start with your worst interaction, fix it, and move to the next. Most sites can get under 200ms with a focused week of work.

Check your own site in 30 seconds

Run a free SEO audit and see exactly where your pages stand on the issues covered in this article.

Run Free SEO Audit

Keep reading