Core Web Vitals after the 2026 INP shift: a practical checklist
By Nimitt Bhatt · 8 February 2026 · 10 min · Updated 20 July 2026
Interaction to Next Paint replaced First Input Delay in March 2024 and quietly turned a lot of green Core Web Vitals dashboards yellow. Two years in, most Indian sites we audit still have not adapted. The pages load fine, the LCP is green, but every real interaction on a mid-range Android is stuttering, and Google now measures that stutter as a ranking signal.
The good news: on almost every site we look at, INP is a discipline problem, not an architecture problem. Two to four weeks of focused work moves most templates from red or amber to green. The bad news is nobody on the team wants to own it, because it lives in the awkward space between marketing, engineering and design.
Why INP is stricter than FID ever was
FID measured only the first interaction on a page, and only the input delay before your JavaScript ran. INP measures every interaction across the page's lifetime and includes the input delay, the processing time and the presentation delay before the next paint. The threshold for a good score is 200 milliseconds at the 75th percentile of real users. On a mid-range Android over 4G in a Tier 2 Indian city, that is not much room to spare.
The pages that fail INP most often are not landing pages. They are product configurators, filter-heavy category pages, dashboard-style shell pages and any template that triggers a heavy React re-render on click. That is where the diagnostic work has to focus.
The diagnostic order that actually works
- →Pull real-user CrUX data first. Lab tools like Lighthouse mislead on INP because they simulate one perfect interaction on a warm cache. Real users are on cold caches, backgrounded tabs and cheap devices.
- →Segment CrUX by page template, not by URL. INP is a template-level problem. Fixing one product page and calling it done is the most common wasted quarter.
- →Inside each template, identify the three interactions users actually do most: add to cart, open filter, submit form. Instrument those specifically with the web-vitals JS library.
- →Profile main-thread work during those interactions in Chrome DevTools, not on page load. The load timeline is a different story that hides the actual INP problem.
- →Defer or chunk the heaviest scripts. Tag managers, analytics, chat widgets and A/B testing tools are the usual culprits. Never lazy-load the script that owns the interaction itself.
- →Re-test on a real mid-tier Android device (a Redmi Note or Realme in the INR 12,000 to 18,000 band) over throttled 4G. Not your MacBook, not an emulator.
The five patterns behind most red INP scores
- →A tag manager firing 40 to 60 tags on every interaction, blocking the main thread for 300 milliseconds while the UI waits.
- →Render-blocking analytics or heatmap scripts loaded synchronously in the head, when they belong in the tail with a defer attribute.
- →Hero images shipped at 3000 or 4000 pixels wide to a 390 pixel mobile viewport, delaying every subsequent paint.
- →React event handlers doing synchronous work: sorting a 2000-item array, formatting currency across the page, or writing to localStorage on every keystroke. All of it belongs in a worker or debounced idle callback.
- →Third-party chat and support widgets that inject 500 KB of JavaScript before the user has done anything. Load them on interaction, never on load.
The two-week fix pattern
On most sites we take on for a technical engagement, the INP work runs in this order. Week one is measurement and prioritisation: instrument the top three templates, capture real interactions, decide which two problems will move the numbers most. Week two is execution: defer scripts, break up long tasks with scheduler.yield or setTimeout, replace synchronous handlers with debounced or worker-based ones, and re-test on a real device. By day fourteen the CrUX curve starts moving. Full recovery in the field takes 28 days because CrUX is a 28-day rolling window.
"Most Core Web Vitals failures are not framework problems. They are discipline problems. Someone has to be paid to care."
What INP does not measure, and why it matters anyway
INP does not directly measure conversion, revenue or bounce. But every study we have run internally on Indian D2C and SaaS sites shows the same pattern: dropping INP from 500 milliseconds to under 200 milliseconds lifts add-to-cart and form-submit rates by 8 to 22 percent depending on the template. That lift is invisible on the CWV dashboard, and it is often the single largest revenue win available to a mature site.
Where this connects
If your INP is amber or red and the engineering team is too busy to look at it, bring in an outside lens. We run a focused technical SEO audit that includes an INP fix plan mapped to your top three templates, or fold it into a longer growth advisory retainer so the fixes stick. For sites over 10,000 URLs, the indexation triage workflow for a 50,000 URL site is the natural companion piece: crawl budget and rendering speed reinforce each other.
Frequently asked questions
- What is a good INP score in 2026?
- Under 200 milliseconds at the 75th percentile of real users, measured over a 28-day CrUX window. 200 to 500 milliseconds is amber and needs attention. Above 500 milliseconds is red and is likely already dragging rankings on interaction-heavy queries.
- Why does my Lighthouse INP look fine but CrUX show red?
- Lighthouse runs one simulated interaction on a warm cache from a fast machine. CrUX measures thousands of real interactions on cold caches, backgrounded tabs and mid-range Android devices. CrUX is the truth. Lighthouse is a suggestion.
- How long before Core Web Vitals fixes show up in Search Console?
- The CrUX window is 28 days, so field data begins moving inside the first week but only stabilises after four weeks of clean measurements. Ranking impact, where present, typically follows another 30 to 60 days later.
- Do INP fixes matter for SEO or only for user experience?
- Both. Google confirms Core Web Vitals as a ranking signal, small but real, and the interaction cliff between 200 and 500 milliseconds shows up in every conversion study we have run on Indian sites. The SEO lift is modest, the conversion lift is not.
Nimitt Bhatt
Founder, SEO Rise. MBA, PGDM in Digital Marketing & Communications, and 20 years across sales and marketing leadership at Reliance Jio, Vodafone and ICICI, now running founder-led SEO advisory across India.
I started SEO Rise in 2024 to work directly with founders and marketing leads, no account managers in between. Every audit, every plan and every reply comes from me.
Book a strategy call
