Yesterday I wrote about using data as a flashlight. There is, of course, a problem inherent in flashlights that is also true about data. If you look at a flashlight the wrong way, you become temporarily blinded. Let someone shine one into your eyes in a dark room and you’ll understand. Data can be blinding too. 
For example, it’s great to have big ears and to listen carefully to what is transpiring with respect to your company or brand in the social sphere. The problem is that we all know those with the loudest mouths tend to be the least satisfied. Some are just chronic complainers; others are trying to get something for nothing. Taking their buzz as gospel can drive you insane as well as point you in the wrong direction. Obviously they can’t be ignored, but that’s a beam of light we need to be sure is aiming in the right direction.
Ratings and reviews are other sources of excellent information, but be sure that as you’re researching (both those of your own brand and those of your competitors) that you’re not falling prey to fake information. There are companies that hire scammers to write them, as this piece explains in detail. Place what’s out there publicly in the context of your own customer service data and support emails. Are there large differences? Complaints that are never made privately but seem to be a steady drumbeat publicly?
The paradigm has historically been to do some qualitative studies to develop hypotheses for testing, then validate and measure through quantitative studies. The only difference now is that, in addition to intimate panel-based research, we also have the ability to get much more input from a panel of millions.
So as you’re using those million beams of light, don’t forget context and source. Make your data set as comprehensive as possible before drawing conclusions. Failing to do so means blindness rather than illumination.
Consult Keith 