WordPress Theme Speed Test: How to Test a Theme Fairly

Why most theme speed tests prove nothing

The typical theme comparison goes like this. Someone installs five themes, imports each vendor's demo, runs one test per theme and publishes a league table. Each demo has different images, different sections and a different set of plugins. The test ran once, from one location, at one moment. What it measured was the demos, not the themes, and mostly the luck of the run.

A fair test has three properties. It holds everything constant except the theme. It repeats each measurement enough times to see the noise. And it records the causes, the files and bytes, not just the score. Get those right and a modest staging site will tell you more than any published shootout. Get them wrong and you will switch themes for a difference that was never there. The WordPress performance hub explains what a theme can and cannot affect; this page is the measuring tape.

If the difference between two themes is smaller than the difference between two runs, there is no difference.

Lab data and field data answer different questions

PageSpeed Insights shows two kinds of numbers on one screen, and mixing them up causes most of the confusion we see. The lab section is a single Lighthouse load on a simulated phone with a throttled connection. The field section, when your site has enough traffic, is a summary of real Chrome visits over the previous 28 days, judged at the 75th percentile.

For comparing themes, the lab is the only tool that works, because a staging copy has no visitors. Field data is the verdict that counts with Google, and it arrives weeks after you ship. The PageSpeed Insights guide for themes walks through each panel of the report.

Build a test bench the theme cannot cheat on

Every variable you leave loose becomes a possible explanation for the result. Lock down as many as you can before the first run.

Content is the hard part. A page built with core blocks renders under any theme, so it travels well between Twenty Twenty-Five, Kadence or Astra. A Divi or Elementor layout does not. To compare builder themes, rebuild the same page by hand in each builder and accept that you are now testing builder plus theme together, which is usually the honest question anyway.

Eight steps from staging copy to verdict

1. 1

   Clone the live site to staging

   Use the same host and PHP version as production. Block search engines on the copy, and note the date of the clone.

2. 2

   Cut plugins to the essentials

   Deactivate everything the test page does not need: analytics, chat, sliders, popups, security scanners. Write the remaining list down. You will add the others back later, one at a time.

3. 3

   Measure a default theme first

   Activate Twenty Twenty-Five and test the page. This is your floor: roughly what WordPress, your host and your content cost before any theme choices.

4. 4

   Build one realistic test page

   Include a header with the real menu, a hero image, a few paragraphs, a form and a footer. Keep it identical across themes, or rebuild it as closely as possible in each builder.

5. 5

   Warm the page before measuring

   Load it twice in a logged-out browser so server and CDN caches are filled, then start the clock. Run a separate cold test if first visits matter to you.

6. 6

   Run the lab test five times per theme

   Use the mobile setting, and test from the same location every time. Keep the median of LCP, CLS and Total Blocking Time, not the best run.

7. 7

   Open the network waterfall

   Count requests, stylesheets, scripts and font files, and total the bytes. Filter by /wp-content/themes/ to see the theme's own share, and check the Coverage tab in Chrome DevTools for unused CSS and JavaScript.

8. 8

   Add the stack back, then watch the field

   Reactivate plugins and the builder one at a time and measure each. After launch, check field data in PageSpeed Insights and Search Console over the next four weeks.

Step three surprises people. If the default theme and your preferred theme come out close, the theme was never the problem, and the time belongs on images, plugins or hosting. Our list of the fastest WordPress themes helps you pick sensible candidates before you start.

The spreadsheet columns worth keeping

Scores are summaries. The causes sit underneath, and they hold steadier between runs. A record like this makes the next test comparable with the last, even months apart, and it gives a developer something to work with if you hand the problem over.

• Context: date and time, theme name and version, builder and version, the plugin list, caching on or off, and the tool, device setting and test location.
• Timings: median LCP, CLS and Total Blocking Time across the runs, plus server response time (TTFB), which tells you whether the wait starts before the theme is involved.
• The LCP element: which element Lighthouse named. If it is the hero image, check it is not lazy-loaded; if it is a heading, fonts and CSS are the likelier delay.
• Weight: total requests, total transferred bytes, and how many of each come from the theme folder, the builder and other plugins.
• Render-blocking files: the stylesheets and scripts Lighthouse flags, with the folder each one lives in.
• Fonts: how many font files load and whether any come from fonts.gstatic.com instead of your own domain.

Keep notes on settings too. When you switch on a theme's local fonts option or its CSS file mode, record it as a separate row. The per-theme speed pages list those switches by name, for example Astra's, Kadence's and Divi's.

Pitfalls that flatter a theme

One more trap is self-inflicted: a theme or plugin lazy-loading the hero image. It makes the lab report look lighter while delaying the largest paint. If LCP is your weak number, read fixing slow LCP. If the layout jumps after load, read theme layout shift.

Reading the result without kidding yourself

Look at the spread before you look at the winner. Say five runs of one theme put LCP somewhere between 2.1 and 2.6 seconds, and another theme's median sits 0.1 seconds lower. That is noise, not a result. A real difference shows up as fewer files, fewer bytes and a different LCP element, and then the timings follow.

If the theme does turn out to be the heavy part, check its speed page before replacing it. Many themes hide the fix behind a toggle: local fonts, CSS file output, critical CSS. If the builder is the weight, a lighter theme will not rescue it. And if speed only fell after an update, start with slow after a theme update. Phones deserve their own pass, covered in theme speed on mobile.

When the bench says the platform is the ceiling

Sometimes the honest outcome of a careful test is that every candidate theme lands in the same place, because the floor is set by PHP, the builder and the plugin stack. That is worth knowing before you spend money on a redesign that will not move the numbers. WordPress vs Next.js speed and Core Web Vitals compared explain why a prerendered Next.js page starts from a lower floor, and where it costs more.

If you would rather not run the test yourself, our WordPress performance consulting does exactly this on a staging copy of your site, and theme speed optimization applies the fixes the test points to. A free site audit is the low-commitment way to start.