You might first try just soil sampling the very high, medium, and low yielding areas of your field as a low level investigative excercise.  Then if you find something interesting from that make your decision to gather more detail and with what approach.

I have found veris to correlate very well to yield on some fields and not at all on others.  The same goes with soil sampling layers and other layers of data. 

The bottom line is everyone talks about zones but just because a data layer is bined out into groups of values (zones) doesn't make it a useable zone to "manage" with.

Instead think of PMA's (profit management areas) and you can't deliniate profit or loss areas without yield data.  Only when you can find data correlating to yield can you then start to have a plan of attack.

Arnie