SEO wizard, Brian Dean from Backlinko, posted a brief video outlining an experiment where he tests whether page loading speed impact SEO ranking.
His goal was to see if drastically improving his page loading speed would move a specific page up in ranking from spot #8. The extremes taken to improve loading speeds were admirable and resulted in a 40x faster page load. And yet, there was absolutely zero impact on SEO ranking for that page.
The page in question was quite competitive, so as he says in the video it's quite possible that improving the page loading speed made some impact, but just not enough.
Other comments in the video mirror Brian's findings, which has totally upended my presumption about page loading speeds. I always assumed that there was a connection. It is useful to see Brian point to stats showing that pages per session and average session duration did in fact go up, but not hugely.