Measuring and analyzing data without having a clear objective is like aiming a bow and arrow without having a pre-defined target.
This means we might just hit something and probably will. But we'll never actually know if it was what we intended in the first place, because we didn't stop to think about it.
Worse, we might just end up making it all about mere opinions instead of objective criteria. If we have a target and we hit it, that's it, we've done it. Otherwise, while we might say we hit something (a tree, an apple, the ground) it's quite easy for someone else to quickly prove we missed a lot of other things.
All it takes is a different opinion to make the whole exercise useless. Objectives matter because they help define what success is all about.
Two ideas about objectives and data that I came across with recently:
1) It's very dangerous to use data exclusively to try and measure human behavior, because numbers themselves aren't the point of measuring things; new insights are. Dan Ariely put it best when he spoke about "trying to understand how a cookie will taste by reading its nutrition label."