As a consultant on Strategic Marketing and Communications, I usually do this for so many 500 Fortune clients my company work with.

Start by demographics (geographics, gender, age, etc) at a macro level. You can dive in it by getting public data that statistics' institutes and companies release periodically.
Once you have narrowed this, go through the next level of segmentation, which is psychographics (interests, behavior, etc) data, usually at a cost, from some typical research companies (or syndicated research). Also from Social Media you can get partial but rich data on this (taking this carefully depending on whether your product lies on digital world or just offline, or a mix of both).
Once you have mapped several segments of population of your interest, start working on the Problem/Solution Fit by mapping which is called "Jobs-to-be-Done"(JBTD) that those consumer segments generically have to accomplish in regards to the problem you're trying to tackle by means of your product or service. A simplistic list of JTBD's can help you narrow and identify the actual problem you can solve. Start iterating then, trying to map the gains and pains you solve for them. These iterations must be triggered after rounds of interviews, either formal or informally, with some representatives of the segments you previously identified. Choose at first some close representatives to you. By this process you will eventually refining your ideal target or targets.

Start small by choosing reduced segments if possible, getting deeper knowledge of them. Afterwards look for scalability by some other bigger segments. It will be the nature of your product what will tell you how small or big you should go at every step.

All in all, it's a hard and dry job, but this way you'll be able to adapt your product and marketing strategy in order to increase the odds to sales conversion once you go to live market.

Hope it helps.