I work at SEOSage, which builds desktop SEO audit software. The steps below work with any crawler.
Google ranks on field data. You are probably measuring lab data.
Core Web Vitals come from real users — the Chrome User Experience Report, 75th percentile, over 28 days. Most audits I see start and end with a Lighthouse run: lab data, one machine, one network, no real users. Lab scores are a diagnostic tool, not the grade. This audit reads the field data first and only touches the lab to diagnose what the field says is broken.
The three vitals, in plain terms
- LCP (Largest Contentful Paint) — how fast the main content appears. Under 2.5s is good. This is the one most sites fail, and it is almost always media and blocking resources.
- INP (Interaction to Next Paint) — how fast the page responds when a user clicks or taps. Under 200ms is good. Heavy JavaScript is the usual culprit.
- CLS (Cumulative Layout Shift) — how much the page jumps around while loading. Under 0.1 is good. Unreserved image and ad slots, late-loading fonts.
You do not need to memorise the thresholds. You need to know which template each failure lives on.
The audit: field first
- Pull the field data. Open PageSpeed Insights for your homepage, one key landing page, and one template page (product, article, category). Then open Search Console's Core Web Vitals report — it shows the same field data grouped by URL pattern, which tells you whether the problem is one page or a whole template.
- Group by template. If every blog post fails CLS, you fix the template once, not fifty pages. Template-level grouping is where this audit saves its time.
- Diagnose in the lab. Only now open Lighthouse — for the failing templates, to find the cause: unoptimized images, render-blocking scripts, font swaps, layout-shifting banners.
- Fix the pattern. Compress and properly size images, defer non-critical JS, reserve space for media and ads, preload the main font. One fix per template, verified in the lab, then rolled out.
Prioritise by traffic
Not all failing templates matter equally. A slow blog archive nobody visits loses to a failing product template that brings revenue. I overlay GA traffic on my audits in SEOSage so the vitals work lands on the templates that earn first — the same traffic-first habit, applied to performance. Any workflow that combines URL-level vitals data with traffic gives you the same ordering.
The field-data lag
Field data moves slowly — 28 days of real users. After you ship a fix, the lab improves immediately and the field report takes weeks. That is normal; do not roll back a working fix because Search Console has not caught up yet. Re-check the field report monthly, and re-run the lab right after every template change — that is when regressions are born.
Disclosure: I work at SEOSage, a desktop SEO audit tool. These steps work with any crawler you already use. When did you last check your field data instead of your Lighthouse score?
Top comments (0)