Establishing a "measurement plane" makes the impedance or admittance displayed on the VNA valid at that specific point. For example, you want to know what the impedance is at the input of some device. You select a transmission line that will reach the device's input and you perform the open,short, load cal with a cal kit of the appropriate connector type for the device under test. Now the impedance measured is the impedance at the device under test's input. Maybe the device under test can be connected directly to the device under test so no cable is needed so the OSL cal is done right at the VNA's test port.
Now if you are simply measuring SWR or return loss the exact location of "measurement plane" is not that critical if its location is not too far down some lossy or poor cable of a different characteristic impedance, etc. (the SWR and the RL remains basically constant along a short transmission line)....Impedance is another thing unless you plot things on a Smith Chart and rotate to the "measurement plane".
Respectfully, I see a lot of discussion about calibration in this discussion group and I think the subject is being overly complicated. The only time I had to get wrapped around the axle is when I had to performance test a VNA in a cal lab. (when I worked at HP/Aglient from '77 to '01, we verified ALL of the published specs and stood behind them so we had to be very careful using the methods and equipment specified by the manual).
With this NanoVNA I think the results have been excellent. I have done a lot of tinkering with impedance matching using it and the measured impedances and the L and C's calculated to perform the matching have worked very nicely.
