Core Web Vitals: does a slow site actually cost you conversions?
Per Google's own 2024 case study aggregations, sites hitting Core Web Vitals thresholds saw 24% lower abandonment. The benchmarks, the trade-offs, and what to fix first for an SMB site.
Core Web Vitals (CWV) are Google's three speed/UX metrics — Largest Contentful Paint (LCP), Interaction to Next Paint (INP), and Cumulative Layout Shift (CLS). They're a Google ranking signal, though a small one, and more importantly they correlate strongly with conversion rate.
What the thresholds are
| Metric | Good | Needs work | Poor |
|---|---|---|---|
| LCP (Largest Contentful Paint) | <2.5s | 2.5–4s | >4s |
| INP (Interaction to Next Paint) | <200ms | 200–500ms | >500ms |
| CLS (Cumulative Layout Shift) | <0.1 | 0.1–0.25 | >0.25 |
Source: web.dev / Google Search Central, current as of 2024
Conversion impact
Per Google's own 2024 case-study aggregations published on web.dev (Vodafone, Carprices, Tokopedia, Renault, others), sites that moved from "poor" to "good" on CWV typically saw 8–24% improvements in conversion or engagement metrics. The largest impact is on mobile, where bandwidth and CPU limits hit harder.
Our own audit data for SMB sites: pages with LCP > 4 seconds convert ~30% lower than pages with LCP < 2.5 seconds when controlling for traffic source and intent.
Source: TNova audits 2025, n=18 SMB sites, mobile sessions only
What to fix first for an SMB site
1. Compress hero images
Biggest LCP killer on most SMB sites: the hero image is a 4MB PNG. Convert to WebP (or AVIF), compress to 60–80% quality, serve responsive sizes via srcset. This single change drops LCP 1.5–3 seconds on most sites we audit.
2. Lazy-load below-the-fold images
Native <img loading="lazy"> is supported in every browser that matters. Add it to every image not in the initial viewport. Bandwidth budget for above-the-fold drops dramatically.
3. Remove unused third-party scripts
The marketing team's drift toward "let's add this tool" leaves SMB sites loading 6–12 third-party scripts: GA, GTM, Meta Pixel, LinkedIn Insight, Hotjar, a chat widget, a popup tool, and three different review-widget vendors. Each script costs LCP and INP. Audit ruthlessly — most can go.
4. Defer non-critical JavaScript
<script defer> or async on anything not needed for first paint. Chat widgets especially — they should never block initial render.
Is CWV worth optimizing if I don't run paid ads?
Yes. The ranking-signal impact is small but real, and the conversion impact is larger. Even if you only rely on organic and word-of-mouth, slow sites cost you the customers who do find you.
Should I switch from WordPress to Next.js for speed?
Probably not just for speed — a well-optimized WordPress site can hit "good" CWV with the right host (Kinsta, WP Engine) and a lean theme. Switch frameworks for editorial workflow or developer-experience reasons, not just for CWV.
What about INP — is it really worse than the old FID metric?
INP is stricter than FID (First Input Delay) and many sites that were "good" on FID are now "needs work" on INP. The biggest fixes: reduce main-thread JavaScript work, break up long tasks, debounce expensive event handlers.
Try the related free tool
/audit →See where you show up when customers search.
Free 24-hour Visibility Audit. Yours to keep — whether you hire us or not.