Most slow sites are slow for the same five reasons — and none of them require a bigger server. Here is the checklist we run through for every site we host.
Run your URL through PageSpeed Insights and note three numbers: LCP (largest paint), INP (interaction delay) and your server response time (TTFB). Fix in that order — a slow hero image beats a slow server 9 times out of 10. Re-measure after every change; guessing is how people "optimize" for a week and change nothing.
Convert hero and content images to AVIF (fallback WebP/JPEG), resize to the actual display width — a 4000px upload shown at 800px is pure waste — and add width/height attributes plus loading="lazy" on below-the-fold images. This single step is usually half the total win.
Text compresses extremely well. Enable gzip (or Brotli) and tell browsers to cache static files for 30 days with immutable fingerprints:
gzip on;
gzip_types text/css application/javascript application/json image/svg+xml;
gzip_min_length 1024;
location ~* \.(jpg|jpeg|png|webp|avif|css|js|woff2)$ {
expires 30d;
add_header Cache-Control "public, immutable";
}
After nginx -t && systemctl reload nginx, repeat visitors barely download anything.
Free Cloudflare plan: DNS + CDN caching for static files + "Always Use HTTPS". Visitors get served from the nearest edge instead of crossing the ocean to your VM. We configure this for every WebDeps site by default.
Use at most two font weights with font-display: swap, preconnect to the font host, and add defer to every non-critical script. Third-party widgets (chat, analytics, embeds) are the usual LCP killers — load them after interaction, not before.
Every WebDeps plan ships with tuned Nginx caching and Cloudflare configured — plus free migration that applies the whole checklist to your current site. See plans →