It depends on the detection method you are using and on the scientific rigor you are applying to the process.
In theory, the only way to determine a true mass percent would be to develop standard curves for the component corresponding to each peak you detect in your chromatogram. This would mean running a pure sample of that component at various concentrations to determine the peak area corresponding to that concentration and preparing a graph (or mathematical function, which is easier these days) of peak area vs. concentration in the sample. If this is done individually for each component of your mixture, then when you inject your mixture, you should be able to read the concentration of each component in the mixture by reference to the standard curve for that component.
In practice, similar compounds have similar properties, and many components are measured by reference to other components rather than standard curves unless you are working in highly regulated lab situations like forensics or pharmaceuticals.