What's a good WordPress load time to aim for?

Under 2 seconds for full page load, under 300ms TTFB. Under 1.5 seconds is excellent. The more useful number is LCP (Largest Contentful Paint) — Google's Core Web Vitals benchmark is under 2.5 seconds. GTmetrix and PageSpeed Insights both show this. Focus on LCP rather than the overall 'score,' which is a composite and can be gamed.

Which caching plugin should I use?

If your host runs LiteSpeed (Hostinger, many others): LiteSpeed Cache. It's free, integrates with the server natively, and handles everything — page cache, object cache, CSS/JS optimization, image WebP conversion. On hosts without LiteSpeed: WP Super Cache is the simplest option, WP Rocket ($59/yr) is the best paid option. Don't stack multiple caching plugins — they conflict.

Does hosting really affect speed that much?

It's the ceiling everything else operates within. A slow server produces slow TTFB regardless of how much you optimize JavaScript. I've seen sites where switching from a $2/mo budget host to Cloudways ($14/mo) cut TTFB from 820ms to 190ms — an improvement no plugin could achieve. Fix the host first; then optimize. Don't optimize around a fundamentally slow server.

Is WP Rocket worth $59/year?

For sites on non-LiteSpeed hosts: yes. WP Rocket is the most complete caching and performance plugin available. It handles page cache, lazy loading, CSS/JS optimization, database cleanup, and CDN integration with minimal configuration. On LiteSpeed hosts, LiteSpeed Cache is free and matches or exceeds WP Rocket. So: if you're on Hostinger or another LiteSpeed host, don't buy WP Rocket.

My PageSpeed score is 45. Where do I start?

In order: (1) Check TTFB first — if it's over 500ms, the server is the problem, not the site. (2) Install a caching plugin and enable it. (3) Run your images through ShortPixel or Imagify, enable WebP. (4) Add Cloudflare free CDN. Run the score again after each step. Most sites at 45 can reach 80+ just from those four changes.

Should I use Cloudflare?

Yes — the free tier is one of the best decisions you can make for a WordPress site. Cloudflare caches static assets (images, CSS, JS) at edge locations globally, adds DDoS protection, and forces HTTPS. Setup is 15 minutes (change nameservers, done). The only caveat: if your site has real-time content like WooCommerce carts or login sessions, configure the caching rules to bypass those pages.

How do I check if my improvements actually worked?

Run GTmetrix before touching anything — screenshot the results (score, load time, TTFB, page size). Then test after each individual change so you can attribute the improvement. Don't make five changes at once and then check. Also test from the same location each time — GTmetrix defaults to Vancouver, which may not match your audience. Pick a location near your readers.

WordPress Speed Optimization: Fix the Right Things in the Right Order (2026)

GTmetrix gives you a score and a list of opportunities. The list is sorted by what the tool can detect — not by what will actually make the biggest difference. Most people end up working backwards, spending time on JavaScript preloading and render-blocking scripts while never fixing the slow server underneath.

Written by Jason Chen · Lead Reviewer

6 years as a web developer. Have optimized WordPress sites from 6-second load times down to under 1.5 seconds. Currently maintain 8 active WordPress installs.

Updated March 17, 2026

Why most speed guides get the order wrong

I ran GTmetrix on a client's site. Score: 68. Load time: 5.2 seconds. PageSpeed flagged 12 separate issues — 'eliminate render-blocking resources,' 'reduce unused CSS,' 'serve images in next-gen formats,' five more. The natural instinct is to work through the list.

Instead, I spent four minutes enabling LiteSpeed Cache and configuring the basic settings. Score: 91. Load time: 1.4 seconds. The render-blocking warning? Gone. The unused CSS warning? Gone. Half the list disappeared because the underlying problem — PHP executing on every request — was now fixed.

The tools surface what they can detect from a browser. They can't tell you your server is fundamentally slow. They can't rank their own recommendations by real-world impact. That's what this article does.

The hierarchy

1. Server speed — the ceiling everything else works within

2. Caching — eliminates PHP execution overhead

3. Images — usually the biggest asset size problem

4. CDN — fixes geographic latency

5-10. Everything else — real but marginal gains

Do 1-4 first. Most sites reach their speed goal before they ever get to 5.

How to diagnose before you optimize

Before touching anything, establish a baseline. Run three GTmetrix tests (results vary; average them) and screenshot everything: score, load time, TTFB, page size, number of requests. You need to know what changed after you make changes.

The single most diagnostic number is <strong>TTFB (Time to First Byte)</strong>. This is how long the server takes to start responding. It's the server's report card, separate from everything else.

TTFB range	What it means	Action
Under 200ms	Server is fast	Move to caching and images
200-500ms	Acceptable, room to improve	Caching may help; consider a faster plan
500-800ms	Server is slow	Server is the problem — no amount of plugin optimization fixes this
Over 800ms	Server is very slow	Switch hosts before doing anything else

If TTFB is over 500ms on a site with less than 30K monthly visitors and no heavy e-commerce: the host is the problem, not your site. Switching hosts will do more than any other optimization.

10 fixes ranked by actual impact

Fix the server firstHighest impactTTFB: -200ms to -800ms

Everything else sits on top of the server. If TTFB (Time to First Byte) is above 400ms, the server is too slow — and no frontend optimization will fix it. Check TTFB in GTmetrix under 'Performance' before doing anything else. Under 200ms is good. Over 600ms and your host is the problem.

Cloudways ($14/mo) cut TTFB on one test site from 820ms to 190ms — bigger improvement than all other changes combined.

Enable a caching pluginHighest impactLoad time: -1s to -3s

Without caching, WordPress runs PHP on every page request — querying the database, building HTML, sending it. Caching builds that HTML once and serves the pre-built file. For most content sites, this is the single biggest performance gain available.

LiteSpeed Cache if your host supports it (free). WP Super Cache for simplicity. WP Rocket if you want everything automated ($59/yr).

Compress and lazy-load imagesHigh impactPage size: -50% to -70%

Images typically account for 60-80% of page weight. Two separate fixes: (a) compression — reduce file size without visible quality loss using ShortPixel or Imagify; (b) lazy loading — don't load images below the fold until the user scrolls there. WordPress has native lazy loading since 5.5; just make sure it's enabled.

WebP format adds another 25-35% compression over JPEG for equivalent quality. ShortPixel converts automatically.

Add Cloudflare CDNHigh impactLoad time for remote visitors: -0.5s to -1.5s

A CDN serves your static files (images, CSS, JS) from edge servers near each visitor rather than from your origin server. If your server is in Dallas and a visitor is in London, a CDN can cut that latency significantly. Cloudflare's free tier covers this and adds DDoS protection on top.

15-minute setup: add site to Cloudflare, point nameservers, done. No code changes required.

Update PHP to 8.2+Medium-High impactServer processing: -15% to -25%

PHP 8.2 is 15-25% faster than PHP 7.4 for WordPress workloads. Yet a large percentage of WordPress sites still run outdated PHP because nobody changed it. Check your host's control panel — it's usually a one-click switch under 'PHP Version.' Test on staging first if you have active plugins, but most well-maintained plugins are compatible.

This is a free performance improvement that takes 2 minutes. There's no reason not to do it.

Audit and remove unused pluginsMedium impactLoad time: -0.2s to -0.8s

Every active plugin loads PHP on every page request — even if you're not using it on that page. Deactivating isn't enough; delete plugins you don't need. Check Query Monitor to see which plugins are most expensive. Common offenders: backup plugins that run database queries on every load, old SEO plugins you replaced, contact form plugins you switched away from.

A site with 8 active plugins loads meaningfully faster than one with 25. Audit with Query Monitor.

Switch to a lightweight themeMedium impactLoad time: -0.3s to -1.5s

Multipurpose themes (Avada, Divi, BeTheme) load 400-800KB of CSS and JS even on simple pages that use 5% of their features. Lightweight themes like Astra, GeneratePress, and Kadence are built to load only what you use. Theme switching requires rebuilding your design, so this is higher effort — but for new sites, start with a fast theme.

One test: switching from Avada to GeneratePress on the same content, 4.2s to 2.1s load time. The difference is real.

Enable Gzip/Brotli compressionMedium impactTransfer size: -60% to -80%

Compression reduces the size of HTML, CSS, and JS files transferred over the network. Most hosts enable this by default — check GTmetrix for a 'compression' warning. If it's not enabled, your caching plugin usually handles it, or you can add it to .htaccess (2 lines for Apache). Check before assuming it's on.

Brotli is newer and compresses better than Gzip; Cloudflare enables it automatically.

Minify CSS and JavaScriptLow-Medium impactTransfer size: -5% to -15%

Minification removes whitespace, comments, and unnecessary characters from code files. Your caching plugin handles this (WP Rocket, LiteSpeed Cache both have minify options). Enable it, test that nothing breaks, and move on. The performance gain is modest on its own, but combined with compression, it adds up.

Sometimes causes visual issues if themes use poorly structured CSS. Test after enabling — easy to revert.

Clean up the databaseLow impactQuery time: -0.1s to -0.3s

WordPress accumulates post revisions, spam comments, transients, and orphaned data over time. On a new site, this doesn't matter. On a 5-year-old site with daily publishing, it adds up. WP-Optimize (free) handles cleanup in one click. Run it monthly. Bigger impact on sites with heavy content creation history.

Backup before running. The cleanup is reversible from backup, not from the plugin.

Before and after: real numbers

Same WordPress site, same content. Starting state: default shared hosting, no caching, unoptimized images, no CDN.

Metric	Before	After	Change
TTFB	820ms	190ms	−77%
Full page load	6.8s	1.4s	−79%
Page size	4.1 MB	780 KB	−81%
HTTP requests	84	22	−74%
LCP	5.4s	1.2s	−78%
PageSpeed (mobile)	29	91	+214%

The TTFB improvement came entirely from switching the host (slow shared → Cloudways). The page size improvement came from image compression + WebP. The load time improvement came from caching. The HTTP request reduction came from caching (combined CSS/JS) and removing unused plugins. The CDN brought further gains for non-local visitors not shown in this test.

Minification, render-blocking fixes, preloading — none of those were needed to get from 6.8s to 1.4s.

Tools worth using

Testing & diagnosis

GTmetrix: Baseline testing, waterfall charts, TTFB measurement. Free tier covers everything you need.

PageSpeed Insights: Google's tool. Uses real Chrome user data. The score your SEO actually depends on.

Query Monitor: WordPress plugin. Shows which plugins add the most load, slow database queries, PHP errors.

Images

ShortPixel: Best all-around image compression. Compresses existing library + new uploads. Converts to WebP/AVIF. 100 free images/month.

Squoosh: Google's browser-based compressor. Good for one-off images before upload.

Caching

LiteSpeed Cache: Free. Best option if your host runs LiteSpeed or OpenLiteSpeed. Handles everything including image optimization.

WP Super Cache: Free. Simplest setup. Good baseline caching without complexity.

WP Rocket: $59/yr. Best paid option. Complete automation for non-LiteSpeed hosts.

CDN

Cloudflare: Free tier is excellent. Add your site, change nameservers. 15-minute setup.

Bunny CDN: Pay-as-you-go from $1/mo. Best performance-per-dollar for high-traffic sites.

Mistakes to avoid

Installing multiple caching plugins

LiteSpeed Cache + W3 Total Cache + Autoptimize running simultaneously creates conflicts. They minify already-minified files, cache over each other's output, and produce broken CSS. Pick one caching plugin.

Optimizing before measuring

Make a change, then measure. Don't make five changes and try to figure out which one helped. You need to know what's working so you can replicate it on other sites.

Chasing the PageSpeed score instead of actual load time

The score is a composite that can be improved without making your site genuinely faster for users. Focus on TTFB and LCP — the two metrics that reflect real user experience.

Skipping PHP version updates out of fear

Most developers test their plugins on current PHP. PHP 8.2 is well-supported. The risk of staying on PHP 7.4 (security vulnerabilities, slower performance) is higher than the risk of upgrading. Test on staging first — it takes 20 minutes.

Not testing mobile separately

Google uses mobile-first indexing. A site that loads in 1.5s on desktop may take 4s on a mobile connection. Always run PageSpeed Insights with both Mobile and Desktop tabs.