Main do's (or don'ts):

  1. Don't pay! There are good free analytics tools. Flurry is just one of them.
  2. Track only the important things. And by that I mean the things that can give you some insight, and that you can act on them.
  3. Tacking onborading users - most important trcking of all. Fiurst experience has a huge effect on retention rates.
  4. A/B testing - don't keep testing the same decision! It will never reach 100%. Set fair goals and try to measure up to them. Then move on.
  5. Forget industry benchmarks. Don't look at others. Just try to succeed in what you do.

Very god article that focused me on what's really important. I highly recommend you stick to these points as well. Full article here:
http://gamasutra.com/blogs/DylanJones/20141120/230576/Indie_Game_Analytics_101_part_2__five_mistakes.php