Thanks for the article. To me these measurements didn't prove as much about Xfce vs KDE, but rather that distributions matter a lot! When the dude's friend tested, clearly you could see that the Arch Xfce was at the level of KDE Neon RAM consumption. So what "bloats" Xfce with Ubuntu is probably Ubuntu itself. Ubuntu with KDE consumed more RAM than Xubuntu. It'd be great to pick the same distribution and try the WMs with that. This way you'd remove the variable what the distribution pulls in, and just then you could come to conclusions to KDE vs Xfce.
Also in a real world system probably you end up with a smorgasbord of KDE, Gnome and other GUI software, which will pull in the KDE, Gnome and other shared libraries. So yet again the distribution bloat will matter a lot, although the WM still matters. The article said for example for the last figure that it's a "close race". However he shows 150MB+ differences in a 1GB scale, that's more than 15%. Not at all close.
Despite all of the sloppiness I'm thankful that the writer at least did some comparisons.