And IMO your take is good! (There's some bad grammar for ya!)
You've correctly described what this really is which is simply a bunch of measurements of lenses in unknown conditions. These are observations only. And it's doubtful the testers have any training in sampling theory to pose accurate conclusions.
Fact here may very be that brand x has a wider manufacturing tolerance than brand z but what does that mean? Was one lens type more expensive than the other since it has a higher manufacturing cost? Does one design have better characteristics aside from the measured clinical sharpness such as pleasing bokeh or fewer artifacts?
I find these types of comparisons being made more for hype and attention when they come from a company which is understandable as this is advertising and publicity to help their sales.
I'll accept the first three paragraphs without argument, actually agree with it basically.
As to the last one, I assume you're thinking I do this to help Lensrentals sales (I don't own it anymore, but a reasonable thought). I would point out that in that case dong m4/3 testing is not a very reasonable use of my time: m4/3 gear is perhaps 4% of their sales and the vast majority of that is video, not photo. Lensrentals hasn't made a suggestion of what I should test in several years. If it mattered to their traffic I'm sure they would at least ask me to test this or that.
Anyway, I basically test whatever interests me when I'm not testing for industry. The main reason for variance testing is we're developing a 'rapid MTF' testing system that requires standards and acceptable variance so I need that data and I share a lot of what I collect. If most people aren't interested in it I can shorten the blog posts a bit, and shorter blog posts are always a good thing.