Speed tests can only tell you so much when not everyone can go the distance.
Website speed is a really important consideration when you’re comparing hosting providers. That’s why we decided to run some tests and see how our plans perform, compared to similar plans at Kinsta and WPEngine. The results are good news for Cloudways customers. This article is a look behind those numbers, into the way we performed the tests between Cloudways, Kinsta and WPengine and what we learned along the way.
We’re reporting results for three different tiers of plans. First, large plans that suit agencies running dozens of client sites. The second tier compares plans for small businesses, and the final tier looks at entry-level plans.
How we decided which plans belong in each tier
Comparing plans isn’t always straightforward, as we discovered when working out which of our plans to test against each Kinsta and WPEngine plan. Here’s a bit about how to compare plans, and why we didn’t match plans by price.
Kinsta and WPEngine plans include a number of monthly visitors. It’s important to realise that a “100,000 visit plan” is a pricing bracket, not a guarantee that every one of those visitors will be served a super-fast website. (This is obvious when you think about it: If all 100,000 visitors turn up at the same time, there’s no way the server will keep up.) Cloudways plans don’t impose caps like this, so we needed to look a little harder.
These visitor caps will become very important later when we get to test results and winners.
The test set-up – same WordPress, ‘standard issue’ servers from different providers
We ran tests with WordPress as the CMS and enough content to replicate the experience of browsing a fairly standard website. The cached:uncached ratio was 60:40. Each website had 10 pages and at least 5 plugins. The servers ran PHP 7.4, with Varnish on, Breeze on, and no CDN.
Hardware
To get the servers up and running we did what you would do – logged into the platform and spun up the plans we were interested in. No tricks, and no changes to what you would get as a new customer.
Independent testing platform: loader.io
We did the testing itself using loader.io, an independent online tool, starting with Cloudways servers. As loader.io simulates traffic it graphs the results. Over time we see the number of (simulated) clients hitting the server, and the server’s average response time.
Screenshot of a genuine Cloudways test run on loader.io
How we defined success and failure: Speed matters
We all know how important it is for websites to be fast. People hate waiting and they’ve quick to abandon sites that take too long. Studies vary but you can bet that after a few seconds a lot of them – maybe most – will leave. SEO experts agree that Google is likely to penalize your site, too.
Setting the bar
“To provide a good user experience, sites should strive to have LCP [largest contentful paint] occur within the first 2.5 seconds of the page starting to load.”
Following Google’s lead, we nominated 2.5 seconds as the threshold for success in our WordPress tests. Google calls this “Core Web Vital” – an important performance signal that the search giant looks for when it ranks websites.
Specifically, we tested for the number of concurrent visitors that each server could handle while still serving the WordPress site within 2.5 seconds.
Calculating results and winners
We tested Cloudways plans first to discover what traffic volume it took to push average load times past the limit of 2.5 seconds. The tests mimicked 5-minute long visits and gradually increased the flow of traffic. The results show us the change in average load time as more visits are received.
By multiplying each traffic volume out to a monthly figure, we see how many monthly visits each Cloudways plan could successfully and quickly serve in a steady stream. Then we compared that number to the visitor caps on the Kinsta and WPEngine plans. You can also check and learn about Bluehost alternative as well.
In every WordPress test, Cloudways blows past the competing visitor caps without any trouble at all, often by a factor of 10 or more. That meant there was no point load testing those more expensive plans, which were guaranteed to lose out on total visits and price.
Results time: Great news for Cloudways customers
The results were good for Cloudways across the board:
- Every Cloudways plan cost under half the price of the competition.
- In every WordPress test, Cloudways plans could quickly serve many times more visitors than the competing Kinsta and WPEngine plans allowed.
