Depending on your desired target, it can be very usefull....
A lot of estimations cannot be carried out in complex matrices or you have over-blending of values false positiv reactions, cannot be viewed in compex mixes or the molecules dont appear in analyzable concentrations etc. the only way to solve this is to clearly separate your molecules, maybe concentrate them and clearly fit togehter the parts of this puzzle
Do you plan to use online-analysis or is a classical back-end sampling ok for your informations?
I think you need to describe the "specific molecule" in more detail. If you cannot disclose the molecule that's perfectly understandable but there are numerous influences on the choice of your method, i.e: molecular weight, functional groups, polarity, charges, pH stability, temperature stability etc... so without specifying your target it will be hard to give a satisfying answer.