WordPress Test Parameters

Disclaimer: Our reviews are not sponsored, as we used our own money to purchase hosting subscriptions. Our reviews are 100% based on our sole opinions. Reviews contain affiliate marketing links where we get a commission when you purchase or subscribe to that service. Read our Affiliate Marketing Policy here. Also, please verify all product and pricing information directly with the linked brands, as product offerings change often. FastestWP.Website is not responsible for errors or product discrepancies.

We proudly host our website on Rocket.net.

We take pride in our WordPress Hosting Reviews, and one of the reasons is that we personally purchase and test each WordPress web host we review. Yes, we used our own money and paid for these web hosting services to test them in real-world conditions. We also test these web hosts using a WordPress site we created with a custom theme, a good amount of images and text, and plug-ins to simulate a person or company’s blog. We don’t just test using the off-the-shelf WordPress theme because not many users will use this.

This page describes our test parameters for the tests we run on each web host. You can decide if you agree with our test parameters and how they may be the same or different from your website.

Our Test WordPress Website

You can see our test website here: Nomading.world. This is our personal travel blog, so feel free to bookmark it! We specifically used this page as the test page for GTMetrix and the Worldwide Load Test. However, we modified the images to be full-size JPEG images, not optimized WebP images as they appear on the page. Our reasoning is that we wanted to test each web host’s image optimization features, and this is only possible with unoptimized images. If we used our own optimized WebP images, we wouldn’t be able to test the ability of each web host to generate optimized WebP images from JPEGs.

Here are the WordPress specifications we used:

• WordPress 6.3.1 (current version as of September 2023)
• PHP 8.2 (current version as of September 2023). Note that some web hosts did not support PHP 8.2 when testing, so we used the most recent version of PHP they did support.
• Match Themes Evoke theme, which includes the MT Evoke plug-in.
• Contact Form 7 plug-in.
• OMGF plug-in (see below regarding our discussion of Google Fonts).
• WPCode Lite plug-in for code snippets. We had to add a code snippet to force our theme to preload the viewport image, as this was not a built-in feature of this theme. Not having this would have destroyed our Largest Contentful Paint (LCP) scores and would not have been fair to the web hosts.
• Yoast SEO plug-in. Yoast is a very popular plug-in we use on all our websites.
• Advanced Custom Fields PRO plug-in came with the theme, so we kept it activated, but we did not use any custom fields on our test web pages.

We also removed any code that required an external server. For example, we removed the Mailchimp newsletter signup as it required fetching a JavaScript file from Amazon S3. While having a Mailchimp (or Constant Contact or another newsletter provider) signup form reflects real-world usage, we felt it was unfair to base a web host’s performance partly on something they cannot control: an external server, such as Amazon or Gravatar.

Additionally, our test website uses two Google Fonts (Roboto and Roboto Condensed), but we used the OMGF plugin to store the Google Font files locally on the server. This increases performance by caching the Google Font files locally and, as explained below, removes any third-party server influence on the performance test by not having to download the font files from Google on each page load. We recommend the OMGF plugin for increased performance if you use Google Fonts.

Caching Plugins and Settings

When we originally designed the test website, we used a hosting provider that used the LiteSpeed Cache plug-in. We used LiteSpeed with any web host that did not use their own or recommend their own caching plug-in or solution. For example, WPX recommended using W3 Total Cache, so we used W3 instead of LiteSpeed when testing WPX.

One item to note is that the YouTube WordPress block uses iframes to deliver and load embedded YouTube videos. We realized during testing that LiteSpeed will lazy load iframes if you turn on this option. However, not all caching solutions and plugins will lazy load iframes. Additionally, while WordPress will lazy load iframes on its own, this is only true if your iframe appears towards the middle and bottom of the page. If your iframe appears towards the top of the page, WordPress will *not* lazy load the iframe. By not lazy loading the iframe – and your YouTube video – your website performance may be severely degraded.

GTMetrix Testing Specifications

We used GTMetrix to test web page performance from 13 locations worldwide. GTMetrix is a well-established web page testing service in the industry, and we trust its results. Here are the specifications we used:

• iPhone 13/13 Pro/13 Pro Max with an LTE connection
• Chrome 103 using Lighthouse 9.6.4
• We loaded this webpage with optimized WebP images.
• We ran GTMetrix twice for each location but only used the scores from the second run. We did this to ensure the page was properly cached on the CDN server from the first run.
• We ran individual tests for 13 locations worldwide:
  • Vancouver, Canada
  • San Francisco, California, USA
  • Chicago, California, USA
  • São Paulo, Brazil
  • Johannesburg, South Africa
  • London, England, UK
  • Frankfurt, Germany
  • Dubai, UAE
  • Mumbai, India
  • Singapore
  • Hong Kong
  • Tokyo, Japan
  • Sydney, Australia

We felt the 13 locations above provided worldwide coverage, with at least one test location per continent (not including Antarctica). We especially wanted two test locations in the United States – one each for the east and west coasts – and we wanted one location for India. We also wanted two test locations in Europe, one in the EU and one in the UK (non-EU).

Note that our test parameters for GTMetrix use an LTE connection, while Google Page Speed uses a 4G connection. We chose LTE for GTMetrix because we felt that Google’s 4G connection is unreasonably slow for many places in the world. Furthermore, although mobile traffic is the majority of Internet traffic, many users on mobile use a WiFi connection, not just a cell connection. Thus, these users will have a faster experience than assuming a 4G connection. However, if you are targeting an audience with a slower connection, you will want to consider this when reading our performance test results.

Worldwide Load Test

Although we tested website performance from 13 locations around the world using GTMetrix, the GTMetrix test only tests one user from one location at one time. This is unrealistic for a website that is receiving web traffic. Thus, we also ran a Worldwide Load Test to simulate 50 users simultaneously using the website for five minutes. We ran this test from 17 global locations using Amazon’s AWS servers:

• London, England, UK
• Ireland
• Frankfurt, Germany
• Milan, Italy
• Stockholm, Sweden
• Israel
• Central Canada
• Virginia, United States
• United Arab Emirates (UAE)
• Mumbai, India
• California, United States
• Cape Town, South Africa
• Singapore
• São Paulo, Brazil
• Hong Kong
• Tokyo, Japan
• Sydney, Australia

Most locations match up with GTMetrix test locations, although we added four additional locations to ensure we had good European and Middle East Coverage.