Hopefully, the procedure you got from a publication will give you some indication of a concentration. If not, you sort of have to pick a concentration and hope for the best. After a while you get good at figuring out which reactions are going to go very fast and may need to be run more dilute, and which ones will be slow and will need to be concentrated. And of course, the solubility of your starting materials or products may determine how much solvent you need.
Usually I would start with ~0.1 to ~1 M of my limiting reagent in the solvent. Smaller reactions I tended to run a little more dilute, just because it is hard to measure and manipulate very small quantities of solvent.