Quick Summary: To rank in order of their effect from greatest to least, the top 5 ways you can optimize your WordPress website for speed in 2026 are:
Server Optimization (via proper hosting and using PHP 8.3) can reduce TTFB by up to 50-70%.
Database Clean-up (the cleaning of SQL queries) reduces query time by 30-50%.
Optimization of an Image Pipeline (via use of AVIF/WebP, Lazy Loading and Responsive Images) cuts Page Weight by 40-70%.
Optimization of Code (by reducing LCP by 500-1500ms via Critical CSS and JS Deferment)
A Full Caching Stack (Page Cache, Object Cache, CDNs and Browser Cache) will provide a full cache solution for load times that are as much as 50-80% faster than non-cached websites.
In the past ten years speed optimization has remained pretty stagnant – “install a cache plugin”, “compress your images” and “use a cdn.” While those suggestions are still valid, the websites that are scoring 90 percent of higher on PageSpeed Insights do a lot more than what is considered basic.
In order of the largest to smallest impact each of these 5 steps build off of one another. Not optimizing for server level will result in reduced benefits from the subsequent optimizations. If you get all 5 right you’ll be able to make sure your wordpress website takes less than 2 seconds to load as well as pass all core web vital thresholds.
Technique 1 — Server-Level Optimization
The server is the starting point. Regardless of how much front-end optimization is applied, no front-end optimizations will overcome a slow server.
The speed of your website’s back end dictates your minimum speed. Shared budget hosting generally produces a time to first byte (TTFB) of 600 ms to 1200 ms. However, managed WordPress hosting by companies such as Cloudways, Kinsta or WP Engine can produce a TTFB of 100 ms to 300 ms. This represents a difference of 500 ms to 900 ms that occurs prior to the loading of an image or execution of one line of JavaScript.
Version of PHP makes significant differences. PHP 8.3 (the recommended version in 2026) will process WordPress workloads approximately 15-25% faster than PHP 7.4. Nevertheless, it is alarming how many WordPress websites continue to run on outdated PHP versions simply due to the fact that someone did not update their server configurations. Check your current PHP version within your host dashboard and upgrade to the latest version if you have fallen behind.
HTTP/3 and QUIC. The latest HTTP protocol reduces connection overhead, especially on mobile networks. Most modern managed hosts support HTTP/3 by default. If your host does not, it is a sign that their infrastructure is not keeping pace.
PHP workers and memory limits. Your hosting plan allocates a specific number of PHP workers — the processes that generate your pages. Too few workers and visitors queue up, waiting for a slot. Ensure your hosting plan provides enough workers for your traffic volume, and set your PHP memory limit to at least 256 MB.
Effect: Improperly configuring server-level optimizations could potentially improve your Time to First Byte (TTFB) from 50-70%, and your site’s overall page load times by 30-40%.
Technique 2 — Database Cleanup and Optimization
WordPress databases get heavy without you knowing it. Each time a post is updated, each transient value created, each unwanted comment left, and each unused options table entry is added into all of those queries that run off every single one of your site’s pages.
Post revisions are the largest contributor to this. WordPress stores every version ever made of every post as defaults. On a website with 200 posts there could be from 2,000-5,000 versions of those posts stored in the database. Cut back revisions for each post down to five (via wp-config.php), then clean up the rest.
Transients are cache values stored temporarily in the database via plugin. Transient cache values created by plugins remain in the database after they’ve been turned off/uninstalled. Thousands of expired transients build up on large websites.
Options that are autoloaded will be loaded into the memory for each page you visit. A lot of plugins store their information in the option table (with autoload set to yes) even though they may rarely need access to this information. If your site has a lot of options set to autoload but do not have to be loaded all the time then you can audit which ones are autoloaded and stop loading those that don’t have to be loaded. This will decrease the amount of memory being used as well as reduce how many queries are run on your site.
Bloat specific to WooCommerce. WooCommerce stores its session data, transient order data, and action scheduler logs in the database. Sites with high volume sales can end up accumulating millions of rows of data in these three tables. It’s important to clean out these tables on a regular basis.
Effect: optimizing your database will usually cut down on query execution time by at least 30% to 50%, and can even improve overall page generation time for bloating sites by as much as 100 ms to 300 ms
Technique 3 — Image Pipeline
Most WordPress web pages have their largest asset in terms of size – images. Optimized homepages typically include images with a total weight of about 300 KB to 800 KB while an unoptimized one may weigh in with around 3 MB to 8 MB.
WebP and AVIF Automated conversion. The new web image formats provide a reduction of up to 50% in file size for similar visual quality when compared to JPEG. WebP is supported by default in WordPress. AVIF is gaining popularity and provides even better compression. Once you upload your image files into WordPress they will automatically be converted to whichever format the user’s web browser supports.
Lazy loading. Below the “fold” or “viewport” images do not need to download until the user scrolls down to the area where the images appear. WordPress has native lazy loading but it is necessary to properly configure it so that the top portion of the page such as hero images does not lazy-load since LCP (the time it takes the page to render) would be negatively impacted.
Responsive Images with Srcset. Uploading an image that is 2000px wide to be viewed on a phone (which has a 400px view port) will waste your bandwidth. Responsive Images create many different sizes of the same image when it is uploaded and then automatically serve up the correct version depending on what size the viewer’s viewing device is. WordPress can do this natively for the “standard” sizes of images; however, for custom implementations you will have to manually configure this.
Cdn Delivery. Using a Content Delivery Network (cdn), to host images, ensures that users will be downloading the images from a cdn located in their geographic area. Also, cdns provide much more aggressive browser caching as well as HTTP/2 and/or HTTP/3 Multiplexing.
The Impact: Typically using a proper image pipeline will reduce a websites total page weight by approximately 40-70%. This type of improvement is likely to be the most significant performance improvement for image heavy web pages.
Technique 4 — Code Optimization
Unused CSS and JavaScript code is a major “silent” performance killer of WordPress websites.
Critical CSS. Move the CSS that is needed for rendering above the fold content out of your stylesheet and into the section of your HTML. The rest will be deferred. This removes all render blocking CSS from your site and greatly reduces LCP and FCP.
Defer JavaScript. Most JavaScript can wait until after the page has rendered, so add defer or async to your tags. This allows the browser to continue parsing your HTML while waiting for your JavaScript files to load and run. In turn this significantly increases INP (Interactive) and LCP (Largest Contentful Paint).
Unloading unused stylesheets/CSS and scripts. all WordPress plugins load assets globally — regardless of page. So a contact form plugin will load it’s CSS and js in blog posts. And a slider plugin will load an animation library on your about page. If you only need a specific script or stylesheet for a single page then removing that asset from other pages can save anywhere from 200 kb to 500 kb of unnecessary download size.
Minification. minifying removes all unneeded whitespace, comments, etc. From CSS and JavaScript files and can shrink file sizes as much as 10-30%. This is probably the simplest thing you can do to optimize your code and something you should always do.
Effectiveness: typically code optimizations improve your LCP (loading content performance) by 500ms to 1500ms and greatly improve your INP (interactivity performance) by reducing how long your browser spends processing scripts on your main thread.
Technique 5 — The Complete Caching Stack
To take the first step toward improving performance with one caching solution, you need to go all the way – develop a full caching stack to eliminate as much redundancy as possible through every level of processing.
Page Caching, saves the fully rendered HTML version of each web page, thus eliminating the need for WordPress to regenerate this content for every visitor. Page caching is likely the largest impact cache layer available. Once a page has been saved via cache, it will be available in milliseconds vs. the time (200ms to 800ms) needed to create this content using PHP and database queries.
Object Caching (using Redis or Memcached), save frequently requested database query results in RAM. Object caching is the second fastest cache layer after page caching. It can potentially cut down on database requests by up to 80%.
CDN Caching, saves static assets (images, css files, js files, fonts etc.) across multiple edge servers around the world. Each visitor downloads static assets from the closest edge server to where they are located instead of directly accessing the assets from your origin server. International visitors may experience significant reductions in asset load times, from seconds to milliseconds.
Browser Caching, informs a user’s browser to hold static assets locally for a specified amount of time. On subsequent visits, when a user loads a previously visited website, they have access to cached versions of the static assets stored on their local machine. Properly implemented cache-control headers can lead to reductions in repeat visit load times of up to 80%.
Effect: An optimally structured caching system can reduce time for a web-page to be loaded from 50 to 80 % in comparison to one that is not cached. When combined with all of the other methods mentioned, it is possible for even slower WordPress sites to achieve loads within two seconds or less.
Measuring Results
You will have no idea whether you are doing things correctly or incorrectly if you don’t measure them. Utilize as many different tools as necessary to verify your work.
Google PageSpeed Insights has two types of test results. One type is lab data, simulated performance. The other is field data, real user experience from the Chrome User Experience Report. For SEO purposes, the field data will be much more important to you than the lab data as that shows users your site works like a site should work. Your goal for all three Core Web Vitals will be to have a Green score.
Waterfall chart found on GTmetrix breaks down where every resource loads, how many requests are made by the browser, and how long every single one took. This tool can help identify areas within your site slowing things down.
Real User Monitoring is tracking how each of your users sees their experience with you as they interact with your site. This can be done through libraries such as web-vitals, or through service providers that are included in many managed hosts. Your lab testing may have shown potential. Real user monitoring will show what it really looks like to your users.
Run your tests from various physical locations using multiple types of computers (desktop/mobile) and check trending over time rather than just running one test.
Want these techniques applied to your site by professionals? Explore our speed optimization service →
This article is part of our WordPress Speed Optimization Guide — the complete resource for maintaining a fast, secure WordPress site.
Frequently Asked Questions
The greatest individual improvement comes from server level optimizations (a combination of your host improving your hardware + your PHP version). However, all of these optimizations work together. The greatest overall improvements come from using as many of them as possible. An example would be a website that uses optimal hosting, optimized databases, image pipeline, optimizes their code, and has the highest level of caching possible, will always be faster than another website that does well on only one.
A professional will need anywhere from four to eight hours for a complete optimization on an average business WordPress website. If you are optimizing a complex WooCommerce website; it may require ten to twenty hours. Once we have optimized your website, there is still much to be done. Your website will continue to slow down over time, so our speed optimization service covers both the first round of optimizations as well as the subsequent maintenance that will keep your site fast.
Professional optimization will test every single change in staging before deploying to production. Aggressive optimization without testing can break functionality. For example; the wrong file being minified (JavaScript), a script that needs to load synchronously being deferred, or caching a page that should be dynamic.
