Earlier I wrote why to use PageSpeed Insights instead of GTmetrix, but this might not be accurate anymore by the end of the month. GTmetrix announced they will be moving to using Core Web Vitals as well.
Perceived performance is describing the aspect of the loading experience towards your visitors. You can have a high pageload time, while users could think the experience is quite alright. Obviously, the other way around is possible as well.
It's not about pagespeed alone
The downside of measuring and displaying the pageload time alone to get a grasp of pagespeed is that it isn't telling the whole user experience story. Heck, even "PageSpeed Insights" isn't the right term anymore, as it is covering technical UX instead of just pagespeed.
Although your site or shop could be quite heavy, maybe even heavier than your online competitors, it could still be possible to have an optimal user engagement by realising optimal First Contentful Paint (FCP), Largest Contentful Paint (LCP) and obviously limited layout shifts (CLS) to not causing any user frustration during the perceived performance process.
If you, as a developer, are able to trigger an early user engagement, you than created some time budget to load other resources in the background. Obviously, you still want things to work as soon as possible to also reduce user frustration in that part of the loading process. However, the mentioned metrics are part of Core Web Vitals and turned out to already reduce the change of early stage page abandonment by 24%. Moreover, Core Web Vitals will become a ranking factor in 2021.
GTmetrix also got hold of this and now decided it is time to change as well.
While [GTmetrix] libraries have served us well in the past, they primarily focused on how well your page is built for speed (or structure), and not how fast it actually loads for your users
GTmetrix.com
How GTmetrix is changing
When doing a (free) GTmetrix analysis, you would see statistics about the amount of resources and the total size of all resources, as well as the fully loaded time. As GTmetrix is stating as well, they were primarily measuring if a website was built for speed (or structure). However, having HTTP/2, the amount or total size of resources aren't likely to be the biggest performance bottlenecks as of 2015. This way of measuring already was outdated for a while.
The way GTmetrix would display pagespeed information before their change.
GTmetrix welcomes Lighthouse and Core Web Vitals for pagespeed analysis
Google experimented with Lighthouse and it's metrics for quite a while. They introduced new metrics, concluded these were telling a good UX story and pushed to production together with Lighthouse v6. It will even become a ranking factor. As a result, GTmetrix couldn't stay behind. This means that the above page loading time and size indicators will be replaced by some Core Web Vitals metrics:
- Largest Contentful Paint (LCP);
- Total Block Time (TBT) while PageSpeed Insights is using First Input Delay (FID);
- Cumulative Layout Shifts (CLS).
The way GTmetrix is displaying pagespeed information as of using Lighthouse and Core Web Vitals.
A new informative panel will be introduced as well, to be able to show speed recommendations on top of structural recommendations.
To use or not to use GTmetrix for pagespeed?
This update means I was both wrong and right. Wrong in the fact that in the near future, which should be at the end of September, you can also use GTmetrix to get a clear image of the perceived performance and thus user experience of your page loads. It turned out I was right in the fact that GTmetrix wasn't indicative towards real user experience.
So yes, at least try out GTMetrix in the near future as well. I know for sure that I can't wait! Happy speeding!