Core Web Vitals 2026: the SEO metric hurting your ranking

Why Google penalises slow sites in 2026

Google confirmed in 2026 that Core Web Vitals remain an active ranking signal. Not the most powerful among hundreds of signals, but the one that tips the SERP when two pieces of content are equivalent. According to the official Google Search Central documentation, a site must pass all three metrics to gain the advantage, with no partial credit.

The logic fits in one sentence: Google does not want to send users to pages that drag. When a competitor publishes an article as good as yours but their site loads in 2 seconds while yours takes 5, Google promotes the competitor. It is mechanical.

Three metrics make up the evaluation, each with a binary good or bad threshold:

These thresholds are measured at the 75th percentile of your real visitors over the last 28 days. If three out of four visitors see your site load in under 2.5 seconds, you pass LCP. Otherwise, Google records it as a negative signal on all similar pages of your domain.

48% of mobile sites fail: which side is yours on?

The 2025 Web Almanac published by HTTP Archive gives the numerical verdict. Only 48% of mobile sites pass all three metrics, and 56% on desktop. In other words, one site in two pays a silent SEO penalty without realising it.

The breakdown per metric reveals where it breaks:

LCP is the weakest link of the web. On mobile, nearly four sites out of ten fail this single metric, which is enough to disqualify them from all three. The cause is almost always the same: a hero image that is too heavy, or a framework that hydrates its JavaScript before showing visible content. On Next.js and Nuxt, the issue concentrated on INP in 2024-2025, before the maintainers shipped hydration fixes.

If your site belongs to the 52% that fail at least one metric, you are mechanically at a disadvantage against a competitor that passes them all. And that disadvantage compounds article by article, page by page.

What LCP, INP and CLS measure (in plain executive terms)

Each of the three metrics describes a precise, measurable, revenue-costing user frustration. No need to memorise the acronyms: remember what they punish in the real life of a visitor.

Each translates into concrete commercial loss. According to benchmarks compiled by Vercel on the impact of Core Web Vitals on SEO, a site that improves its LCP by 1 second gains on average 10% in conversion rate, and a site that passes the thresholds on all three metrics sees its organic traffic grow by 15 to 25% in 6 to 9 months.

The "100 on PageSpeed" trap: why Google does not look at it

Most executives open PageSpeed Insights, see a Lighthouse score of 95, and conclude that everything is fine. That is exactly the wrong reasoning. Google does not use this score to rank. It looks at the CrUX (Chrome User Experience Report), which aggregates real data from millions of actual Chrome visitors.

The distinction is crucial:

The Lighthouse vs CrUX gap explains why so many sites with a "good PageSpeed score" stall in Search Console: Lighthouse is a lab test, CrUX is the field audit. Google ranks exclusively on the second one.

How to audit your site in 5 minutes without a dev

You need zero technical skill to know whether your site passes or fails Core Web Vitals. Four free tools, in order, give a reliable answer in under 10 minutes:

1. 1

   PageSpeed Insights, Origin Summary section

   2 min

   Paste your site URL on pagespeed.web.dev. Below the Lighthouse score, scroll down to "Discover what your real users are experiencing". That is the CrUX field data. Green = you pass. Orange or red = you fail.
2. 2

   Google Search Console, Core Web Vitals report

   3 min

   In GSC, "Experience" tab then "Core Web Vitals". The report lists exactly which URLs fail, by group and by metric. It is the official Google source.
3. 3

   CrUX Vis or CrUX Dashboard

   2 min

   CrUX Vis (Google tool) displays your metrics evolution over 25 months. Lets you see if you are regressing or improving since your last release.
4. 4

   Vercel Speed Insights or equivalent RUM

   2 min

   If your site is on Vercel, the Speed Insights tab is free up to a certain volume and gives detail per page. Otherwise: DebugBear, SpeedCurve, or the native Web Vitals integrated via the official useReportWebVitals hook.

If the assessment is bad, you will know exactly which metric fails, on which pages, and at which percentile. That is the minimum information to brief a provider.

The levers that bring a site back under the thresholds

Four families of technical actions solve 80% of Core Web Vitals issues. They are not exotic operations, but they need a developer who knows where to look.

None of these levers requires rebuilding your site. They are surgical optimisations that a competent developer handles in 2 to 5 days, depending on the starting state. For the broader 2026 SEO/GEO challenges beyond Core Web Vitals alone, our dedicated SEO and GEO pillar frames what comes next.

What to ask your technical provider

The most common risk in a poorly scoped CWV audit: a provider who comes back with a Lighthouse score of 100 and declares the job done. The right questions to ask before commissioning the work:

A good provider will answer all five clearly. A bad one will say "we will optimise performance" with no precise commitment. The fifth answer is worth knowing: Google takes at least 28 days to recompute its CrUX report, and 2 to 3 additional months for the SEO benefit to show in rankings. Any provider promising an effect in 2 weeks is wrong or lying.

The ROI of a Core Web Vitals audit

On a site doing 5,000 organic visits per month, unlocking Core Web Vitals typically represents 30 to 60 K€ in additional annual revenue in B2B services. The math is simple: a site that moves from "Poor" to "Good" rating sees its organic traffic grow 15 to 25% over 6 to 9 months, and its mobile conversion rate increase 8 to 12%.

For a site already well ranked on its main queries, the gain is more diluted but still measurable: we typically observe a 2 to 5 position improvement on secondary queries, where CWV play their tie-breaker role against equivalent competitors. That is exactly what we observed on the refactor projects we delivered in 2025-2026, and it is what our Performance maintenance pack covers with the monthly CWV audit + CrUX tracking + 4 SEO articles per month.

On the Odyssée home, we went from LCP 18.5s to 9.4s in a single refactor session. The gain cascaded to all child pages within 2 months, and Google indexation accelerated in the wake. The trajectory is nothing exceptional: it is what any well-run CWV audit produces.