As the co-founder and CEO of (probably) the world’s fastest WordPress hosting service, often WPX customers will come to us being very alarmed about a seemingly bad ‘Performance Grade’ or ‘Performance Score’ on tools like Google Page Speed Insights, Pingdom Tools and GTMetrix.
And while these diagnostic tools do have value, I am now going to present the case as to why you should COMPLETELY ignore these figures and what to focus on instead.
EXHIBIT A: YouTube's Performance Scores
We all know YouTube.com, the video sharing site that Google acquired way back in 2006.
It’s a phenomenon of course and currently – according to Alexa – is the SECOND most popular site on the planet:
YouTube and its parent, Google, have virtually unlimited resources to do anything with the site, from any technical angle.
Plus, and you can check this with a free Tunnelbear VPN, YouTube.com loads INSTANTLY from almost anywhere in the world.
You would imagine therefore, that the site has been incredibly well optimized for [a] massive traffic, and [b] loading speed.
So why therefore, would YouTube.com’s desktop score be a lousy 52 out of 100 and a ‘Slow’ rating on Google’s own speed testing tool:
and on Mobile, YouTube.com didn’t do much better with a score of 55 out of 100 and a ‘Moderate’ speed rating:
In the interests of fairness, when you run Google’s Page Speed Insights through itself (highly recommended on any speed testing tool), it did score a perfect 100 on Desktop (94 on Mobile) BUT that only deserved a ‘Moderate’ speed rating?
And the ‘accuracy’ doesn’t get much better when we analyze YouTube.com on Pingdom Tools (San Francisco test location) where the score is a ‘C’, 73 out of 100 and a reported 2.75 seconds loading time (that time is also nonsense as I discuss here):
Things get even more bizarre over on GTMetrix.com with a reported load time of over 10 seconds, an ‘E’ ‘Page Speed’ score of 57 out of 100 and a YSlow score of ‘C’, 70 out of 100 – remember this is for YouTube.com, the second most popular site in the world that loads instantly from pretty much anywhere! Are you starting to see a pattern here?
If you are worried about performance grade type scores, I strongly urge you to run the same tests yourself that I have here.
Let’s move on to more evidence AGAINST these time-wasting Performance Grades shall we?
EXHIBIT B: Apple.com's Performance Scores
So the world’s first ever trillion-dollar company and the inventor of the iPhone can only muster a ‘pathetic’ 31 out of 100 for Mobile but speed rated at ‘Moderate’ on Google’s Page Speed Insights? On Desktop, they earned an 87.
On Pingdom Tools (San Francisco), Apple.com could only manage a grade of ‘C’ (79):
While On GTMetrix, Apple.com ‘struggled’ with a ‘PageSpeed Score’ of ‘C’ (73%) and ‘YSlow Score’ of ‘C’ (74%) + a loading time of 2.7 seconds (though, again, it loads instantly from anywhere in the world):
I doubt if Apple CEO Tim Cook is too stressed about Apple.com’s terrible performance grade scores and in my view, neither should you when it comes to your site.
FINAL EXHIBIT: Their Performance Scores On Themselves?
We already saw that Google’s Page Speed Insights did at least give itself a high, but not perfect on mobile, performance grade but what about Pingdom Tools?
So Pingdom Tools managed a ‘B’ score of 88/100 here which is not awful but not amazing either.
But do Pingdom take their own advice on Performance Grades?
Pingdom Tools lives on a subdomain at tools.pingdom.com but what about their main root domain where the serious online business is done at pingdom.com?
Ah, apparently not. It only scored a ‘D’ with 62/100.
I guess Pingdom themselves don’t regard their own Performance Grade measurement as important or they would follow their own advice – on their main income generating site!
For the record, GTMetrix came in with a ‘A’ in both categories for their own tool with 99% and 90% scores:
CLOSING ARGUMENT: What To Look At INSTEAD Of Performance Grade Scores?
As I discuss here, even page loading times on these tools can be misleading but they are useful in identifying possible bottlenecks for real human visitors, not bots e.g. UNoptimized images, slow-loading themes/templates.
Generally, and we saw this above with YouTube and Apple, real, HUMAN visitors, not bots from Pingdom or GTMetrix or Google Page Speed Insights, will see the page MUCH faster than the REPORTED time shown with these tools.
The easiest test is to go to your own website/s (on WPX, for example) from different VPN locations (Tunnelbear has a useful free VPN option) and browsers with cleared caches and see how fast they load for you, a real person visiting the site.
Alternate browsers include Microsoft Edge, Opera and Brave.
Case closed (hopefully!).