I have heard from one or two people that Google's Page Speed Tool is not important. Is that right? Is it OK if my page is 16/100 on Google's tool but gets excellent scores on GTmetrix? Are tools like WebPageTest and Pingdom also not important? Thanks
[filtered from 38 💬🗨]
It absolutely matters, and will matter more next year when Google bakes the metrics into their algorithms. Page speed and speed perception are crucial for better conversion rates.
thanks, I know speed matters. It was specifically about whether I can ignore scores in the teens on Google’s tool because the site gets good ratings on GTmetrix.
One or two people selling page speed services have said it doesn’t matter what Google’s Tool says, as long as GTmetrix says it is good. Just doesn’t feel right.
Andrew » Mat
CSS-PURGE ~ Helping you burn that extra CSS off
Mat ✍️ » Andrew
thanks for the great information. We have a developer who can custom code websites but he doesn’t know how to speed them up, or maybe he is too busy on other stuff to learn it. I think we will pay someone to do it as I am pretty clueless about code and whatnot, but I will certainly read your link and take on board your other info. Cheers 👍
Andrew » Mat
no problem. Have your guy check out https://web.dev too as loads of resources on speeding sites up on there.
It absolutely doesn't matter at the moment. In the future – maybe it will.
GTmetrix is not very good. Have a look at this https://web.dev/vitals/
Thanks Kevin. This page doesn't mention GTmetrix. Please can you explain why it is not very good? Is it that GTmetrix doesn't address Core Web Vitals? Is Google's Tool the main one I should be looking at?
Makes sense to me that Google's Tool should be the most important as they make the rules for their algorithm. But we have been told to ignore it by people who sell speed optimization as a service.
Kevin » Mat
the page was for an explanation of what google considers. I should have been more clear. Apologies.
Re GTmetrix, we ran several comparison tests across 4 tools and GTmetrix gave overly optimistic scores on site speed and several core web vital metrix.
We use 3 tools Page Speed Insights as an initial diagnostic tool, and Lighthouse plus WebPageTest.org for deeper analysis.
This is what’s worked for us over the past 5 years.
I think people who say use one tool only are wrong. I also think that people who say ignore the Google tools are also wrong.
There are also additional tools used by dev teams to refactor and improve code besides testing server speeds and so on.
Awesome information Kevin, thanks for sharing.
I know about Core Web Vitals and have been explaining it to clients. It has just put a spanner in the works that at least one of the services we tried said to forget about Google's Tool and just look at GTmetrix, when I had already pointed out client's slow speeds using Google's Tool i.e. they are expecting an increase in the score if we get someone to speed up their site. Will try another service.
Kevin » Mat
you are most welcome. If I find more information I’ll post and share 😊
Kevin » Mat
Core Web Vitals: What They Are & How to Improve Them
Kevin » Mat
and https://support.google.com/webmasters/answer/9205520?hl=en. Be aware that Google sets targets
Core Web Vitals report – Search Console Help
I often put off reading his articles as they are long 😜but will read it now. And the 2nd you also added. Thanks 👍
Think about it. Google’s not basing their rankings on GTmetrix. Google’s data always trumps others when it comes to rankings. That being said, page speed is not a major ranking factor. You should be more concerned with user experience on that and not a page score.
That is exactly what I think too. But we have been trying speed optimization services and they say different.
We used two services. One has a page loading in 10 seconds on GTmetrix with A scores (but it loads faster in reality). Then on Google's tool the page scores 16/100. The other service got Google's scores in the 90s, but it broke various things on the website and I think we had to basically undo what they did. But I believe they were saying something similar about ignoring Google's Tool.
Eric » Mat
yeah, there are a number of factors that will affect page speed scores… where the server is that’s pinging you, what they use as a benchmark, etc… But no bot can substitute for user experience. If you serve up your desktop site in 2ish seconds and mobile site in 3ish, you should be able to focus your efforts on other more important factors without a seeing any significant negative impact from page speed.
Obviously, the faster the better. Especially for eCommerce sites.
Google Site Speed absolutely matters.
While GTmetrix is great for seeing loading speed, the ratings don't mean much.
Forget about page scores, Focus on page load time, Aim for 2 seconds or less… UX no layout shift or minimal… that's all…
That sounds logical, at least for now.
On GTmetrix the page load time varies and is 10 seconds at the moment, for fully loaded. But it gets very good scores for everything else.
On Google's Tool it gets 23/100 at the moment (was 16 earlier) and the Largest Contentful Paint (not sure if this is the best metric to use) is 21.7 seconds.
Because the page loads fast on incognito mode on desktop (seems fairly instant but I think desktop isn't as important) and gets good GTmetrix scores, we are being told it is fast. For me, it's more like 5 seconds on mobile, but I don't know if I have a slow internet connection and whether I should be trusting the tools more than just my phone. Confusing. Just doesn't feel right at all to me that the site definitely takes longer than 3 seconds on my phone and has terrible scores on Google's Tool.
Would you say Largest Contentful Paint is the most important metric on Google's Tool?
Lime » Mat
Yes and No, In a way how I look and evaluate sites is how fast they load, I use google dev tools in the network section to evaluate what's loading I am surprised why your website is taking 21 seconds to load. When I design and make sites I try to have them fully loaded in the view port less than 2 seconds and UX. Imagine typing google.com and you have to wait until 21 seconds for the website to show or even 5 seconds because naturally how we are all used to browsing after typing url we want to see the website… that's it…
Please note, I am not saying what google page speed or GTmetrix says is wrong, By any means if you can implement all the suggestions they give you like deferring java script serving http2 or serving scaled images optimize images etc but your primary goal should be to look at fully load time to be less than 2 seconds instead of what score you achieve.
So something that google recommends you use on how to fix problems and speed issues if a waste of time 🤔 I think not
G's tool is indeed a good one. It shows recommendations which "needs" improvement, but that does not mean the page "should" be tweaked for improvement. I suppose 16 is a very low score. Upon anecdotal evidence, if your site loads in < 3.0 secs, cool. Else, check it's recommendations and tune that best suits your requirement.