Mostly when we say RF ripple, we mean an unflatness of response vs. frequency.
This is very often caused by mismatch between measurement equipment and the DUT, with the ripple frequency inversely propotional to the cable length between instrument and DUT.
The effect on the measurement is to make the DUT appear to have more freuquency response than it does.
In most cases, it can be removed with post correction, if the source, DUT S11 and S22 and load match are known. Alternatively, using high quality attenuators can reduce the mismatch appear from the source or the load. It is best to add the attenuators onto the ends of the cables, as cables are a source of mismtach as well.
Ripple can also be caused by bad RF cables, bad flatness response in the source, bad flatness response in the receiver.
If you are using an SA to measure amplitude, the YTF can be non-repeatable, and can be a source of mismatch error.
For the most part, any measurments done on a DUT to measure gain or flatness would be better done with a network analyzer, provided the stimulus and response are appropriate (eg, won't work too well on an optical to RF transducer).