Yield
 
Yield is the most important consideration for profitability. The 'compare varieties' at the University of Wisconsin can be useful, but only if there are a considerable number of tests. I suggest at least 4 good tests.  See how well the variety does in comparison to the test mean. This doesn't work well with  leafhopper varieties since they usually appear in tests only for leafhopper varieties, making comparison with other varieties difficult.   Usually the only way to get a significant number of comparisons is compare varieties with the Vernal, the check variety used in most Midwest and Eastern trials. I usually use all locations and all years to obtain the most data. Data from all locations and all years can be used as companies obviously aren't going to place their varieties at locations where they are not adapted.   Since alfalfa is a perennial, it must not only perform under various summer conditions, but also winter extremes. This is why I consider data over many locations to be as important or to backup years in nearby locations.  Most Alfalfa varieties are adapted over wide geographic locations
 
Solution: The Internet allows you to 'compare varieties' as well as look up performance in your state and surrounding states in individual trials. If longevity, three or more full hay years, is important look for varieties with good stand ratings.  If you are looking at individual test data look at the CV (Cofficient of Variation), and LSD (Least Significant Difference). A CV of less than 10% indicate a good test, the data becomes questionable with a CV over 15% or an MCV (Modified CV) over 10%.  CV's on stand ratings and other characteristics usually run much higher; here rely on the LSD. Don't put much faith in yield data from companies selling the variety.