Servizio · Performance & Core Web Vitals
Performance & Core Web Vitals
A site that loads in 6-8 seconds isn't a slow site, it's an invisible site. Since May 2021 Google uses Core Web Vitals as a ranking signal: if LCP, CLS and INP don't pass the threshold, you fall behind tidier competitors.
On mobile, half the users leave before seeing the first heading. The problem isn't "buy a more expensive host": it's understanding where the site wastes milliseconds and cutting.
4MB hero loaded as PNG, fonts loaded without preload, render-blocking JavaScript, layout shifts caused by a badly placed cookie banner, images served without optimization.
No magic: I measure with Lighthouse + WebPageTest + CrUX, identify real bottlenecks, apply fixes, re-measure. If hosting, CDN, cache layer or build pipeline need touching, I do it. The Google check passes or it doesn't close. Period.
01 — Awareness
A slow site is an invisible site
Google penalizes sites that fail the Core Web Vitals check. Mobile is where everything happens: if the first paint arrives in 4 seconds, users leave.
- 01
Hero taking 6 seconds
Cover image served as 3-4MB PNG, without fetchpriority, behind a cookie banner that blocks render. LCP out of threshold, first visitors already bounced before seeing the title.
- 02
Layout jumping while you read
Cookie banner without reserved height, hero without explicit dimensions, fonts changing weight mid-render. CLS over 0.25, the user clicks where they didn't want and the site loses trust immediately.
- 03
Clicks that respond after 1 second
Blocking JavaScript, hydration in a single block, listeners causing layout reflow. INP over 500ms. The user presses "Add to cart" twice because the first click looks ignored.
02 — What you get
6 features- 01
Lighthouse + WebPageTest + CrUX audit
Measurement on 3 cross-checking sources. Lighthouse for lab data, WebPageTest for filmstrip and waterfall, CrUX for real user data over the last 28 days. Only after that we touch code.
- 02
LCP < 2.5s
Largest Contentful Paint below threshold. Hero images served in modern format (AVIF/WebP), display font preload, fetchpriority on the cover, render-blocking CSS eliminated. First the content shows, then everything else.
- 03
CLS < 0.1
No annoying layout shift. Explicit dimensions on every image and iframe, height reserved for cookie banner and sticky header, font-display swap without jumps. The site stops "jumping" while loading.
- 04
INP < 200ms
Interaction to Next Paint below threshold. Long tasks broken with yield, heavy events moved to requestIdleCallback, optimized hydration. Clicks respond immediately, not after waiting for an 800KB JS bundle.
- 05
Image + asset pipeline
Images re-encoded in AVIF/WebP with fallback, native lazy loading, responsive srcset. Font preload only for above-the-fold weights. JS bundle analyzed and tree-shaken without dead dependencies. No "secret optimizations": measurable and replicable.
- 06
Before/after report
Lighthouse + WebPageTest snapshot before and after, with waterfalls compared and numbers on key metrics. If the delta isn't visible in data, the work isn't done. No decorative screenshots: numbers.
03 — How I work
5 phases- 01
Discovery call
30 free minutes to understand the real need. I listen, ask questions, take notes. No cold quotes: I need to know what you're actually building first.
- 02
Quote
Flat fee with clear scope, timeline and costs. No surprises: if something falls out of scope I tell you upfront, not when the invoice lands.
- 03
Design / Strategy
Wireframes, moodboard or audit with action plan — depends on the service. You see the direction before any code or content gets touched.
- 04
Build
Development, technical work or operational execution. Agreed check-ins along the way, none of that "let's sync end of month".
- 05
Launch & follow-up
Go-live + 30 days of assistance included. Documentation, training if needed, and an open door for the small things later.
04 — POST-AUDIT
Performance isn't a one-shot fix
The Core Web Vitals check is green today. The next deploy arrives with another heavy image, a new tracking plugin, an embedded video added on the fly. Without continuous monitoring, metrics degrade in months. That's why after the audit I deliver an operational checklist: what to check before publishing every new page, how to optimize images autonomously, when to re-audit. For sites with significant traffic or critical e-commerce, I activate quarterly monitoring: PageSpeed Insights API + alerts on real regressions. No smoky dashboards: just the numbers needed to intervene before Google sees them.
05 - What you get
Deliverables
| No. | Deliverable | Format | Timing |
|---|---|---|---|
| 01 | 5-10 page baseline audit | PDF report + JSON | week 1 |
| 02 | Bottleneck map per metric | Notion / Markdown | week 1 |
| 03 | Image pipeline (AVIF/WebP + lazy) | Code + CI script | week 2 |
| 04 | Optimized font loading | Code + preload | week 2 |
| 05 | JavaScript splitting + tree-shake | Bundle analyzer report | week 2-3 |
| 06 | Layout shift fix (CLS) | Code | week 3 |
| 07 | INP optimization (yield + idle) | Code | week 3 |
| 08 | Re-audit + waterfall comparison | PDF before/after | week 3-4 |
| 09 | Maintenance checklist | Markdown | pre-handoff |
| 10 | Continuous monitoring setup (optional) | PageSpeed API + alert | post-handoff |
07 - Related services
- 01
Web Design
A fast site is born from design, not chased afterward. Performance enters at the wireframe stage.
- 02
Web Development
Modern stack, smart hydration, tuned build pipeline. Performance isn't a layer applied later.
- 03
Website maintenance
Continuous monitoring to catch regressions before Google sees them.
08 - Free audit · 20 minutes
Find out if your site passes the Core Web Vitals check
I measure LCP, CLS, INP on the 3 most important pages of your site and send a report with the 3 most impactful fixes. No sales pitch: just numbers.
No obligation - reply within 24h
09 — Frequently asked questions
6 answersThe questions I hear all the time.
How long does it take to pass the Core Web Vitals check?
Depends on how badly the starting site is set up. A modern site with some targeted fixes gets sorted in 1-2 weeks. A legacy site with monolithic jQuery, outdated WordPress plugins and 4MB PNGs takes longer, because hosting or build pipeline often need touching. The first audit clarifies the real size of the work.
Can I do it without changing hosting?
Almost always yes. Most problems are in code and assets, not the server. But if hosting is shared and TTFB stays above 1 second even after cache + CDN, the only way is to change it. Promising "miraculous performance" on €3/month shared hosting is worthless.
My site is on WordPress: does anything change?
No, the metrics are the same. The toolkit changes: caching plugin (W3 Total Cache, WP Rocket), image optimizer, WP_Query optimization, possible migration to WP-tuned hosting. For more serious cases, I flag headless migration or alternative CMS, but it's rare.
Do you guarantee the check always passes?
No, and whoever guarantees it lies. I guarantee applying measurable best practices and documenting every fix with before/after. On very large sites or with constraints (third parties, heavy tracking, video embeds) some metrics can stay borderline. In that case I say so upfront, not at project end.
What happens if I change the site after optimization?
Without maintenance, metrics degrade within months: new plugins, unoptimized photos, added tracking scripts. That's why I deliver an operational checklist and, if needed, a quarterly monitoring retainer.
Do you only work on sito-v3 / Next.js or other stacks too?
I work on any stack: WordPress, Shopify, Webflow, custom React/Vue/Next, static Hugo/Astro. The principle doesn't change: I measure, identify, fix, re-measure. The toolkit changes, the metrics stay LCP, CLS, INP.
10 - Start here?
Ready to start?
A 30-minute call to figure out what's actually needed. No PowerPoint.