Can you just do a spatial join instead (using a join merge rule in the field mapping) then use field calculator or something to sort out the most common occurence after?
