For a one-man shop, trying to be a build-to-print machine shop in the classic sense is a tough battle. You might be able to make it work if you're lucky enough to find a couple of big sugar daddy customers. But that's very risky. Here in the 21st century, it's so easy to find alternate sources and work over long distances. For a big company purchasing agent who's under heat to save every penny, there's very little loyalty. If they can find some sucker who'll do it cheaper for acceptable quality, they'll jump suppliers in a minute. That's just the way the world is now.
In my opinion, there are several keys to building up a successful one-man business:
First, become really, really specialized. Find a narrow niche and build up your expertise and reputation in that area, so that you become the "only" guy to go to in the US when someone needs that product or service. Concentrate your business around a narrow product, rather than trying to be the "I can build anything you need" guy. The product can be a packaged service, for example, rebuilding starter motors for early Jeeps. Or offering special ruggedized starter motors for early Jeeps. It should be something that's easy to describe and define. The point is that you want to become the world's expert on early Jeep starter motors. That narrow group of people who need one will search you out.
Second, forget the local market and concentrate on the whole US. There are probably only a few people in your town who need an early Jeep starter motor, but in the whole US, there are enough to keep you busy for many years. Don't waste your effort with a sign out front or an ad in the local Yellow Pages. List a couple of your super starter motors on ebay, and get the word out on some Jeep forums. You want your business to be almost all mail order.
Third, the optimal product for a one man shop is something small enough that it's easy to ship cross country, and is in a price range of, say, $200 to $1000 each. Yes, you can build a business selling $10 products by mail order, but you have to think about how many you'd have to sell per month, and how much effort it would be to deal with that many individual customers and shipments. A $200 product is about the minimum that's worth the effort to transact the sale and ship the product. A $500 product that ships in a small box is even better. A $10,000 product gets dangerous for a one-man shop. The gap between sales gets longer and more uncertain, and the risk is much higher if there's a screw-up or delay. Ideally, you want a product of complexity/cost that you can build and sell 10 per month. That spreads out the risk, but keeps enough flow going that you get some manufacturing efficiency.
Anyway, these are just some things to think about.