One important figure is the amount of time it was found on top of the stack in the cumulative profile summary. Just because a function is found on the stack does not mean that it uses up much processing time, but if it is found consistently on the top of the stack then it is likely that this function has a significant execution time. Another thing to check is that you expect the functions near to top of the call tree to take significant time.
It must be remembered that the numbers produced are from random samples and thus it is important to be careful in interpreting their meaning. The rate of sampling is always coarse in comparison to the function call rate and so it is possible for strange effects to occur and significant events to be missed. For example, "resonance" may occur when an event always occurs between regular sampling times, though in practice this does not appear to be a problem.