---
title: Core Web Vitals
category: product
entity_type: skill
price: $15
canonical: https://forgehouse.ai/skills/core-web-vitals/
lang: en
hreflang_alt: https://forgehouse.ai/tr/skiller/core-web-vitals/
last_updated: 2026-06-20
---

# Core Web Vitals

> Measure, diagnose and optimize Core Web Vitals (LCP, INP, CLS) for web applications using…

A complete audit and optimization workflow for Core Web Vitals: LCP, INP, and CLS: that measures real performance, diagnoses the bottlenecks, and produces a prioritized fix list. It combines lab data from PageSpeed Insights and Lighthouse with field data from the Chrome User Experience Report, so decisions are based on what real users actually experience. The output is a JSON health score plus an opportunity-ranked action plan.

## Use cases
- Capturing a performance baseline before a new site or page goes live
- Fixing URLs flagged in the Search Console Core Web Vitals report
- Diagnosing slow LCP, sluggish INP, or unexpected layout shift
- Running a pre-deploy regression check on page speed
- Resolving a lab-passes-but-field-fails discrepancy using real-user data
- Auditing multiple template types: homepage, listing, detail, blog, search

## Benefits
- Pages hit the Good thresholds that influence Google ranking and Page Experience
- Faster load and interaction reduce bounce and lift conversion
- Fix effort goes to the highest-impact opportunities first, not guesswork
- Performance is verified against real users, not just synthetic test runs

## What’s included
- PageSpeed Insights and CrUX query recipes for lab and field measurement
- The Good, Needs-Improvement, and Poor threshold table at the P75 percentile
- Defensive optimization patterns for image priority, reserved space, and long-task breaking
- Font-loading and third-party-script deferral fixes that prevent shift and INP stalls
- A multi-URL per-page audit loop with opportunity ranking
- A gotchas and edge-cases guide plus a full verification checklist

## Who it’s for
Web developers, SEO specialists, and site owners who need to measure, diagnose, and fix Core Web Vitals with real-user data backing every decision.

## How it runs
Lab numbers flatter, field data ranks. The loop pulls both from PageSpeed and CrUX, lets real-user P75 values win every disagreement, and ranks fixes by estimated millisecond savings across every template type.
1. Calls the PageSpeed Insights API for mobile and desktop and pulls the lab LCP, INP and CLS values plus the Lighthouse performance score.
2. Queries CrUX for the field data, the real-user P75 values Google actually ranks on, and resolves any lab-versus-field disagreement in favor of field data.
3. Extracts the failing audits and ranks the fix opportunities by estimated millisecond savings, so the top 5 fixes are chosen by impact, not by guesswork.
4. Diagnoses each metric at its source: render-blocking resources and hero image priority for LCP, long JavaScript tasks and third-party scripts for INP, unsized images and font swaps for CLS.
5. Repeats the measurement across template types (homepage, listing, detail, blog), because a fast homepage routinely hides a slow product template.
6. Issues a PASS, NEEDS_REVIEW or FAIL verdict against the official thresholds (LCP 2.5s, INP 200ms, CLS 0.1) and re-measures after fixes against the saved baseline so improvement is proven, not claimed.

## FAQ
### My site is new with little traffic. Does it still work without field data?
It can capture a lab baseline from PageSpeed and Lighthouse before launch, which is exactly the no-traffic case. Field data from CrUX joins once real users arrive to confirm what they actually experience.

### Is this just rerunning Lighthouse for me?
It goes further by pairing that lab data with real-user field data, because a lab score can look fine while CrUX shows users hitting slow LCP. The output is a diagnosed bottleneck list, not a single number.

### Does it apply the fixes or only tell me what is wrong?
It measures, diagnoses and hands you a prioritized fix list ordered by impact, but implementing each change in your code is your step. The skill tells you what to fix first and why, not how your stack should ship it.

## Price
$15, one-time, no subscription. VAT included.

Related guide: [How to automate SEO and AEO with Claude](https://forgehouse.ai/guides/automate-seo-claude/)