Often, some will wonder why Googles PageSpeed Insights displays a much better score than a 3rd-party tool or speed checker like GT Metrix. For publishers displaying ads, or any site with media or page elements below the fold (the non-viewable portion of a page upon the page load, Ezoic lazy loads these elements to ensure a fast speed (also known as asynchronous loading). This means that the ads, images, videos, or other rich elements meaning only rendered to the page when the user scrolls to the point where the ad location or media is visible on the screen.
This form of loading frees up bandwidth to load in all the important elements on your site first, so your users are able to engage with your content quicker. It's a practice that's actually recommended in Google's webmaster guidelines and is considered a best practice by ad ops professionals and developers all over the globe. However, GTMetrix isn't designed to render a page the way the user does (PageSpeed Insights uses Google Lighthouse and actual Chrome user data from the browser). GT Metrix and other tools won't be able to distinguish between an element that's been lazy-loaded and one that isn't, which leads to artificial inflation of page load times. While this may not be true for every element ion the site, it is true for many (especially ads).
This is true for both Ezoic users that are using Leap and also those just using Ezoic for other features, like Monetization.
Many of the recommendations given by tools like GTMetrix are simply a checklist of things to look for in a 'perfect' site, without taking into account any context whatsoever. In many cases, the items on the GTMetrix checklist are not even related to site speed or performance. You can see here for example that The Wall Street Journal's site gets terrible scores for this:
Ultimately, it is the user experience that matters more than a site speed score; however, Google's PageSpeed Insights metrics are the ones Google uses as a factor for ranking your site, not the GT Metrix scores.