The Great American WordPress Hosting Speed Test, pt. 1: Welcome

written by: Andrew Martin | October 2, 2014

00004

I’ve been working with, and thinking about, web hosting a lot lately. My last post, which offered general tips on choosing a web host, made me want to dig deeper and put some of the more popular web hosting options to the test. Great reliability, and a responsive support team are both incredibly important but tricky to test objectively in a limited timeframe; I would measure an appropriate sample size for evaluating a host’s reliability in years, which doesn’t lend itself well to being a blog topic. Even with mediocre web hosts, problems arise sparingly, and, as such, conducting a head-to-head challenge of web hosting support departments (without manufacturing your own issues, which I’m not about to waste support agents’ time doing) proves difficult.

One key facet of a quality web hosting option is relatively simple to put to the test in an objective, timely manner: speed. We could simply sign up for a few web hosting packages, upload an identical site, and run Google Page Speed on the home page, but we’re going to take it a bit further. After all, “speed” doesn’t only refer to an end user visiting your home page, it also includes:

• Setup time - how long does it take to sign up for hosting, upload your site, and configure it?
• Front end operations - how fast is your site to a front end user doing normal browsing?
• Back end operations - how long does it take to complete common administration tasks on the back end?
• Backup restoration - something went wrong - how long before you’re back up and running?

The Contenders

Most of sites we build for clients use the WordPress platform. Therefore, this test will focus on hosting packages specifically marketed towards WordPress sites. Does one need a “WordPress hosting” package to successfully host a WordPress site? Absolutely not. In fact, this site (itself a WordPress site) is hosted on a non-Wordpress-specific server, and we’re quite happy with the performance of it. Upon completion of this blog series, I plan on testing several “general” web hosting packages using the exact procedure to see if WordPress-tailored hosting really does offer any speed advantages.

WordPress-specific hosting does offer certain advantages outside the realm of speed, a prime example being support teams who are often more able to offer assistance with WordPress issues, rather than simply saying that the application is the responsibility of the customer, not the web hosting company. This extra service level does cost a bit more than the typical “budget web host”, however; $30 per month is perhaps the de facto standard for WordPress-specific hosting plans, with most companies offering a plan at or near that price. For this reason, we’ll test the $30 per month hosting plans from several leading WordPress hosting providers:

Test 1: Setup

The first test measures how long it takes to get from square one (not even having an account with a hosting provider) to having a fully functional, “typical” WordPress site. I use quotes because WordPress sites vary in size from single splash pages to extremely popular blogs with thousands of posts. I wanted to test using a site that was slightly more complicated than the default post-installation site with the twenty-fourteen theme, one post, and one page. For this reason, the installation I performed on each site included a free theme from a commercial theme provider and four popular free plugins. Although we use some premium plugins on client sites, and we build our own themes from scratch, the decision was made to use all freely available components in case a reader of this blog would like to perform the test themselves. This checklist was followed click-for-click on each of the five tests.

I practiced the entire setup procedure a few times on a development server in attempt to ensure I didn’t get faster for the later tests. I tried to maintain a quick but not frantic pace throughout. As you can see, the results were fairly consistent throughout, with the main (subjective) observable differences coming from differing checkout UX (I found Flywheel’s to be excellent) and a relatively variable duration when importing the sample data (I suspect it was mostly downloading the imported images from remote servers; WP Engine took a fair bit of time to complete the import). Of course, this is a non-scientific test as we have no way of controlling the performance of a remote server at any given time. The bottom line of test 1 is, however, I didn’t have to wait long to get started with any of the five hosts.

Coming Up…

In part 2, we’ll get into the nitty gritty of the testing. We’ll try several common front end operations in various scenarios to see if any of the hosting options excel above the others. Later, we’ll break out our stopwatches and see how long the ever important backup restore takes on each host, as well as test basic back end administration speed.

» Part 2 has been posted. Click to read it!