I’d say those measurements are about as good as you might hope. 0.2% is only about 1/2 a millimeter of error over the full 256mm width and depth of the print bed.
Prints warp, even PLA (just not as much) because plastic changes size as it changes temperature. A few percent, even for the best plastics. More for something like ABS. And print beds aren’t always as flat and well-trammed as they should be, and even if they are the components change size/shape as they’re heated in use. Along with other structural elements of the printer, none of which the printer compensates for even though they’re changing constantly as the machine goes through thermal cycles. Those errors, I should think, are going to dominate (and could be the sole source of the errors you measured with this test).
The way I deal with this, where I need very tight dimensional tolerances on some area(s) of a print is, I print it with those surfaces slightly over-dimensioned and then I run a pass with my CNC milling machine to bring them down to the tolerance I require. It’s a huge PITA. Mostly I just design with looser tolerances to make allowances for my printers. None of them are as accurate as what I can produce with a CNC.