When I start on something, I usually think about all of the components it would actually need, in order to succeed. A 'half-baked' business plan, so to speak. Sometimes, I start a project in order to test the entire plan, but usually, I start a project either to test just one aspect of it, or, as often as not, just to scratch an itch. There should always be an idea of the "growth engine" that could potentially drive the business. (Which may change.) If you want it to be a business, don't think of it as a product, think of the entire business. Though, you don't have to implement the entire business right away. Find out, as quickly as possible, if there's any chance of viability, whether you'd actually like to spend more time on it, and whether some of your assumptions are true or false. A project isn't a business.