When you are talking about "flat" ground I think it is absolutely necessary to use RTK and for that matter on any type of ground.  On flat you can probably throw out the tilt factor BUT on "steep" ground perhaps not.

When you look at the +- specs of RTK vs other corrections and then take the rough rule that the vertical is 2 to 3 times worse than the horizontal it pretty much rules out everthing but RTK.  Another way to look at this is for grins I was given a bunch of data for a field to be analyzed.  There was WAAS elevation data from a harvest layer and another set of data from a WAAS Veris layer.  The Veris WAAS elevation did a better job of discribing undulations than did the yield WAAS data.  Tilt compensated might have been part of it but I think the other part is it took days to gather vs the Veris in hours.  So the Veris activity had less time to express the WAAS error vs the combine activity.  While that would be less of an issue with "4-inch" differential systems in the end RTK year to year is a lot tighter and thus during the data collection as well.  Now take than on done to a smaller event and then you might consider "mission planning" to decide the best time to collect the data.  For those reasons often times it is best to specifically collect elevation data rather than getting it with other data.

Now one could strike up a good debate about using lasers and/or total stations simply because of their better spec in the vertical dimension.  However they have their problems as well but in the end the "flatter" the ground the better you want the vertical spec within reason.  And we haven't even talked about the machine control issue of fixing the problem.  It has to be done with machine control because the human eye is just not good enough and why go to all that effort and then eye ball the dirt moving event.

RTK for elevation analysis may well pay for itself as quick or quicker than tractor driving once the romance of tractor driving wears off.  So I would first worry more about getting the data collected with RTK and then ponder the issue of terrain compensation after that.

Arnie