Google Speedtest Feels Wrong

Hey guys,

What’s your experience with Google Speedtest?
I have warnings and claims that the website takes over 4 seconds to respond (First Contentful Paint), but… it doesn’t.

Am I crazy?

Going crazy…

Optimized Images → Check
Lazy loading → Check
Replace Images with WebP → Check
Minify CSS ->Check
Minify JS → Check
Minify HTML → Check
Load Js Asynchronously → Check
Load CSS Asynchronously → Check
Load external scripts via Google Tags (all possible and some that shouldnt) → Check
Cache → Check
Cloudflare → Check


Try using Pingdom.

Yeah, I am, along with GTmetrix and Google Speedtest.

It’s killing me…
A lot of the complaining is about external stuff, like Google Fonts, Analytics, Hotjar…

Gogle speedtst wants me to allow the page to load without jquery… which according to it it would speed up the site in 0.33

All external scripts, includig grom google…

da fuck…

What is the site/url you are testing?

PM sent , thank you! :kissing_heart:

@404error maybe give a try?

Looks like an extended version of Google Speedtest…

I honestly don’t care about some arbitrary page speed test. It’s just a gimmick.

Problem is when you’re running ads and Google ads starts showing some fk’ed up score (it now shows up on Google ads metrics for each ad), which means that I get some hate coming my way…

What do you mean? In this world with jQuery preloading 4MB of libraries and a 8192x41960 background image, shouldn’t we try to do the best we can!?

Biggest issue I’ve found is caching of external items - google will penalise you for using google analytics because of their own cache policy.

There are ways to skip GA if it’s GT metrix/PageSpeed accessing the site, alternatively you can load in GA after page load but GT will still account for this. The other option is to locally host a copy of GA.

Anything like google fonts, always host locally so you can set your own caching values.

Normally, as @deank says, I wouldn’t bother with what the value says as long as the site feels fast to users - but Google does use it’s PageSpeed rank in terms of how you appear in results - a faster site will get a higher score on the same keyword (so I’m lead to believe from a SEO guy I work with - we spent a long time getting their site from low 40’s on mobile, to 96 mobile/99 desktop).

Happy to have a look if you PM site URL?

It’s telling me to remove all CSS and scripts from the web page in order to increase performance :rofl:. Well I guess it’s true…

if ($is_pagespeed) {
    echo 'Hello, World';
// rest of site

Best PageSpeed score…


That’s still too slow. This is 1 nanosecond faster which means you’ll rank 60/100 instead of 40/100:

if ($is_pagespeed) {
    die('Hello, World');
// rest of site

I did consider that way when posting.

Clearly too early for fully optimised code

1 Like

Just a blank page is the way to go. Who needs content nowadays especially when you are designing a site for bots.

1 Like

I haven’t seen this to be the case (same with jQuery) if you appropriately use async defer.

You forgot to send proper headers tho…