In general, for chemical analysis, the bigger the sample, the better. After all, to determine what an unknown substance is, you have to expose it to certain known reagents and see what, if anything, happens; and some of the sample gets destroyed each time, whether or not you discover anything new about it. You can minimise wastage by careful choice of test sequence (HEY, SPROUT!!! THIS IS YOUR TERRITORY!) but there will always be some.
If you are really unlucky, you can run out of the substance under test before you even find out what is in it, let alone how strong it was .....
Also, the accuracy with which the concentration of a sample can be determined depends upon the size of the sample. You need to know how much powder reacts with a known amount of some other reagent; this will tell you how many molecules of the active substance must have been present in the powder; then, with some further calculations, the percentage by weight. You may already have a rough idea from earlier tests, but you need to do a precise test, adding tiny amounts each time. The "give or take" amount is plus or minus half the last increment that completely annihilated the known substance; and the smallest amount you can add is usually fixed by your apparatus. The smaller any errors as a proportion of the quantity, the better will be the result.
