Hi http-archive team!
I read in your FAQ that you test the webpages on your private instance of WebPageTest. Then I tried out the public webpagetest tool with some pages of which I know that they use lazy loading on images and found out that webpagetest doesn’t recognize this and only counts the bytes of the images “above the fold”.
So, how about your results? Do they include lazy loaded images?
I found this statistic: State of Images
It states that currently 10% of webpages use lazyloading, with a large rize since July 2020.
In case your stats do not include those images, the figures (e.g. median page weight) might get biased if more pages use lazy loading.
It measures everything that was actually loaded in the loading of the page. If a site is using lazy loading and those images aren’t in the viewport, they aren’t loaded (so they aren’t measured).
It’s not biased - it’s accurate for what sites load on the initial load.
When a user actually interacts with a site then the behavior changes. Scrolling something like twitter which is an infinite scroll will load content forever. Same applies to lazy-loaded images. It delivers the same experience and performance for long-form pages.
