Performance

The Website Performance Check Guide

Performance auditing is only useful if it produces a prioritized fix list. This guide walks through the metrics that matter, the order to test them in, and the optimizations that produce real-world improvements.

The metrics that matter

TTFB (Time To First Byte) — how fast your server starts responding. Server/network problem if high.
LCP (Largest Contentful Paint) — when the main content appears. Front-end problem if high.
FID / INP — input responsiveness. Usually JavaScript blocking the main thread.
CLS (Cumulative Layout Shift) — content jumping around. Reserve space for images and embeds.
Total page weight — under 1.5MB is healthy; over 3MB hurts mobile.

Optimization priorities (in order)

Enable a CDN — often the single biggest improvement.
Compress images (WebP/AVIF) and serve responsive sizes.
Defer or async non-critical JavaScript.
Add HTTP caching headers and a server cache (Redis/Varnish).
Eliminate render-blocking resources in the <head>.
Remove unused JS (tree-shaking, code splitting).
Self-host or async-load third-party tags (chat, analytics).
Preconnect to critical third-party origins.

How to interpret your numbers

If TTFB is high but page render is fast → server/hosting work. If TTFB is fast but LCP is slow → front-end work. If both are slow → start with the server (a slow backend makes front-end fixes invisible). After every change, re-measure from the same conditions to confirm improvement.

Recommended next steps

Diagnostic workflow

1
Baseline uptime + response time
Confirm the site responds reliably before chasing milliseconds.
Uptime Check
2
Run a full Speed Test
Measure download, upload, and latency from your real connection.
Speed Test
3
Measure ping & jitter
Network instability shows up as inconsistent loads.
Run Ping
4
Verify DNS isn't the bottleneck
Slow DNS adds 100–500 ms before the first byte.
DNS Speed Test
5
Set up performance monitoring
One-off tests aren't enough — monitor continuously and alert on regressions.
Configure monitoring

Frequently asked questions

What's the most important performance metric?

Largest Contentful Paint (LCP). It's what users actually perceive as 'loaded' and what Google scores. Aim for under 2.5 seconds on mobile.

What's a good TTFB?

Under 500 ms is good; under 200 ms is excellent. High TTFB almost always points to server-side issues — slow database queries, missing caching, or far-away hosting.

Should I test on mobile or desktop?

Both, but prioritize mobile. Most traffic is mobile, and mobile networks are slower — what looks fine on desktop may be unusable on a 4G connection.

How often should I run performance checks?

Continuously via synthetic monitoring, plus a manual deep audit after every major release. Performance regressions sneak in with every deploy.

Does CDN really make that much of a difference?

Yes — usually 100–300ms reduction in latency for users far from your origin server, plus offloading static asset bandwidth. It's the highest-impact, lowest-effort improvement for most sites.