My first companies were all services companies so they immediately generated reveneu and were thus easily bootstrapped. I believe it was easier begin 90s to 2000s to do that though. With that money and also during that period I started building products: a dating site, a CMS, a freehosting OS and large freehoster, a POS for grand cafe's and some niche LOB dev products. I sold all of those after running them for a while; in the end they were all very profitable and fun because of the diversity. The company formed around the CMS grew to 300 people in the high times. That I did not like at all so I quit as CTO and opened the R&D dep of that company with 2 colleagues.
Currently I'm working on the hardware and software for a smart credit card while I have a services company to offload work to and do work for other clients.
Services still work well as bootstrap; first do some freelance work in a client office, start discussing working from home when the trust has been built, start suggesting to get more people on board when it gets busy. Before you know it you have a small team and you can start cloning that process until you have enough portfolio to do less intensive sales. You will never run losses: quite the opposite. When your team takes over enough you can work on side projects if you want to.