To me that's a vague question. However, let's assume you have an idea for a trading strategy using an indicator.
Your first task would be back-testing: go back through historical data, looking for "set-ups" which suit the strategy you are thinking of using, and see how profitable each set-up is; you want to identify 100-200 cases but more is better as you would be identifying set-ups through different economic climates - what may have worked in the last 5 months may not have worked in the last week with the Greece problems.
The next task is forward-testing, so on a demo account, take trades when your set-up occurs and see how profitable each trade. Again, you want to be taking 100-200 trades but more is better.
Finally, forward-test with a live account risking small amounts for 100-200 trades. This introduces the psychological aspect of trading. You could skip stage 2 as many choose to but don't be tempted to risk normal sizes per trade after just back-testing as you will be in for a shock.