Test Granularity

An optical illusion. Square A is exactly the same shade of grey as square B. See demonstration. (Photo credit: Wikipedia)

In testing assumptions, I consider it very tricky determining just how granular I will need data. I could throw every in bit.ly and see what sticks or just focus on the key metrics:
- Sales
- Follows
- Participation.

I think I can manage with just those, but if I can't resolve any assumptions, then that will be the trigger that I need more detailed data.