There's really no more to it than I've said. I use a long coax cable to connect a reference dipole antenna a few wavelengths away to one VNA connector and the antenna to the other. It usually is easiest to connect the antenna under test to S1 so you can observe the match as well.
For a reference dipole I use a tiny rabbit ears that came with my RTL-SDR because it can be adjusted to be resonant at most VHF/UHF frequencies (it came with two sets of elements). It also has a ferrite isolator installed on the cable. Make sure it's oriented correctly and away from other objects and correctly adjusted for the frequency by adjusting the elements for maximum signal at the frequency of interest.
My extension cables are made of RG316 and were bought premade from Amazon. The loss is much lower than RG174.
I use microphone stands, stage lighting stands and camera tripods to support the antennas (whatever works) - sometimes I use small sections of PVC water pipe to extend them to keep the antennas away from the stands.
The results are excellent out-of-doors, but indoors I have to do some trial and error to find measurement locations for both antennas not affected by reflections, but usually find one fairly easily.
The results are only broadly useful, but I compare the performance with another reference dipole in the same location and it's usually good enough to save a lot of time before going to play in the back yard.
Make sure, of course to calibrate the S1 port at the antenna itself first. I usually do a local (short) cable calibration "through", but I probably should experiment with calibrating out the remote antenna. For my purposes, manual comparison with reference is usually enough...
M
