You want your WordPress site to load lightning-fast. And if you’re like most of us, when you think of improving your site’s page load times to get that “lightning-fast” designation, you think of your Google PageSpeed Insights score.
For many website owners, it’s their white whale. Getting a perfect score on PageSpeed Insights is the impossible quest that will magically solve all of their page speed woes.
But is a high PageSpeed Insights score the be-all and end-all of fast page load times? Sorry, but no. If your focus is on improving your site’s page load times, finding a better host will often take you further.
In this post, I’m going to run a real test to show you that high-performance hosting will do more for your page load times than endlessly striving to improve your PageSpeed Insights score.
What is Google PageSpeed Insights? Should you care?
If you’re not already familiar, PageSpeed Insights is a Google-offered tool that helps you both analyze and optimize your website’s performance for desktop and mobile visitors. Before I get into what exactly that entails, let’s talk about what PageSpeed Insights is not:
It’s not a tool that gives you a real number for how long your page takes to load. You need other tools for that. Instead, PageSpeed Insights gives you a score from 0-100 based on how well your site is technically optimized, with separate scores for desktop and mobile.
To calculate this score, Google uses two metrics (direct quote):
- time to above-the-fold load: Elapsed time from the moment a user requests a new page and to the moment the above-the-fold content is rendered by the browser.
- time to full page load: Elapsed time from the moment a user requests a new page to the moment the page is fully rendered by the browser.
And while Google’s suggestions are important, they readily admit that the test “only considers the network-independent aspects of page performance”. That means PageSpeed Insights does not, among other aspects, directly consider the speed of your host.
So while PageSpeed Insights gives you great suggestions for speeding up the technical aspects of your page, getting a perfect score on PageSpeed Insights will never be enough to completely overcome a slow host.
In Google’s own words, the “absolute performance of the page will still be dependent upon a user’s network connection.”
* It should be noted that Google is also talking about a visitor’s personal Internet speed.
Do visitors care if you score 100/100 on PageSpeed Insights?
No. Most of your visitors won’t have a single clue what PageSpeed Insights is, let alone care about your site’s score.
Plenty of data says that users care about how quickly your page loads in absolute terms. Akamai found that, at least for eCommerce sites, 30% of users want sites to load in under 1 second. And on mobile, 74% of users will abandon a site that takes longer than 5 seconds to load.
(Charts by Visualizer Lite.)
But as we learned in the previous section, PageSpeed Insights does not calculate the absolute page load time of your site.
Does a high PageSpeed Insights score equal a quick-loading site?
Yes and no. A higher PageSpeed Insights score will likely improve the speed of your website. It will get you a quicker loading site.
But here’s the thing:
Your website will only ever be as fast as the hardware that’s powering it. A high-quality web host and good caching will almost always take you further than a 25-point improvement in your PageSpeed Insights score.
To prove that, I’m going to run a real-world test…
Here’s how I’ll set it up:
- Two fresh WordPress installs. One on Kinsta (high performance managed hosting) and one on Arvixe (cheap shared hosting from the same parent company as Bluehost and HostGator).
- Default Twenty Seventeen theme
- Jetpack installed and activated (no Photon CDN, though)
- Imported WordPress test data
- One post with one ~150 KB full-size unoptimized image
Using those settings, I’ll run each site through PageSpeed Insights and use Pingdom Tools (testing from San Jose) to calculate the absolute page load time for each site. I’ll use PageSeed Insights’ Desktop score for the purposes of this test.
Once I have those numbers, I’ll optimize both test sites to improve their PageSpeed Insights scores by:
- Installing W3 Total Cache (only on Arvixe, not Kinsta)
- Installing Autoptimize for better minification
- Optimizing the image
My goal is to get at least a 90+ PageSpeed Insights score.
Once I achieve that goal, I’ll run both sites through PageSpeed Insights and Pingdom again. And then, I’ll show you all of the data!
**NOTE** – Kinsta has built-in caching. So when I first run the Kinsta site through PageSpeed Insights, it will naturally have a better first score than Arvixe. Therefore, it’s most illustrative to compare the optimized versions of each site against one another.
Test results – PageSpeed Insights vs. moving to faster hosting
After running the tests and crunching the numbers, here are the results I came up with.
Arvixe Test 1 – Unoptimized PageSpeed Insights score:
My unoptimized Arvixe test site had a mediocre PageSpeed Insights score of 66/100:
Pingdom’s calculated page load time was better than I expected, clocking in at 1.95 seconds. That’s actually not horrible. But remember that I’m testing on a dummy site without much content:
Kinsta Test 1 – Unoptimized PageSpeed Insights score:
With the exact same settings as Arvixe test #1, my Kinsta test site had a much-improved PageSpeed Insights score of 80. As I mentioned in the disclaimer above, this is because Kinsta has server-side caching:
As I expected, Kinsta also had a superior page load time of just 610 ms:
Arvixe Test 2 – Optimized PageSpeed Insights score:
After making the tweaks I mentioned above, I got the PageSpeed Insights score for my Arvixe test site up to 90/100. A major improvement from 66 and worthy of the coveted “green” seal of approval:
So what was the effect on actual page load times? A marked improvement to 947 ms. Which illustrates my point…as significant as that improvement is, it’s still slower than the unoptimized test site on Kinsta:
Kinsta Test 2 – Optimized PageSpeed Insights score:
After making the same tweaks on Kinsta to optimize images and minify code, I managed to get its PageSpeed Insights score up to 91/100, almost identical to my Arvixe test site. The only difference is that Google dinged the Arvixe test site 1 point for “Reduce server response time”:
Pingdom results weren’t much different. There was a slight improvement to 574 ms, but nothing mind-blowing:
You can view a summary of the results below. Remember – these are identical test sites. Literally, the only difference is that I used W3 Total Cache for my Arvixe test site, while Kinsta already included built-in caching.
|Host / Test #||PageSpeed Insights Score||Pingdom Page Load Time|
|Arvixe Test 1||66||1.95 s|
|Kinsta Test 1||80||610 ms|
|Arvixe Test 2||90||947 ms|
|Kinsta Test 2||91||574 ms|
Wrapping things up
I don’t want you to come away from this post thinking that you should ignore Google PageSpeed Insights. The people at Google are way smarter than I am. They know what they’re doing.
All I’m saying is this…
Focus most on what’s actually important – absolute page load times. PageSpeed Insights is not the end of your optimization journey. Nor should it always be your primary focus.
When it comes to absolute page load times (the important thing!), switching to a faster host and implementing good caching will have a much larger effect than trying to go from 66 to a 90 PageSpeed Insights score on sub-par hosting.
By all means, implement PageSpeed Insights’ suggestions. Just don’t obsess over them.
What do you think? Are you struggling to make your WordPress site load fast?