Not always. The idea behind this method is to eliminate all but 1 variable by assigning values for s.
To determine the value of A, you made s = -1, which eliminated B and C, which gave you the value of A.
Now that you have the value of A, it's no longer a variable; it's now a constant, which you can substitute back in your equation.
By assigning s = 0, you eliminated B, leaving you only C to worry about. The idea is to keep assigning values to s in order to eliminate variables until you have one variable left (in this case you found A and C, leaving you only B), and now you have to assign a value which would "get rid" of s, and leave only B and constants.
You could also try songoku's method, which would give you the same answer.
http://www.exampleproblems.com/wiki/index.php/Algebra-Partial_Fraction_Decomposition has a lot of examples, and step by step solutions (with the method you used).
Also, there are some PFD's that you might need to know some matrix row reduction to solve, so you might want to get a refresher on that if you need it.
Good luck with Differential Equations. Had to do that last semester; it was a pain.