If looking for a previously isolated compound, on can just follow the experimental listed in the paper that described the compound. If looking for a novel compound, one can use assay directed fractionation. I like to use a column screen to determine the best column and potential mobile phase to purify the active compounds. After the column screen, I can optimize both the extraction and chromatography based on the compound behavior on the different columns.
Some people like to do sequential extractions of their plant mass, starting with hexane or petroleum ether, followed by ethyl acetate, then dichloromethane, methanol, and water. This resolves the compounds somewhat by polarity, but the same active compound(s) may be in multiple extracts. My experience is that methanol or ethanol extraction is a good start, and one can optimize the extraction with the rest of the plant mass after the compound polarity is determined.