I am a patent agent, and have gotten patents for several small inventors. There has been some good information in this thread, especially from G-dog, and an important point is that whether you need a patent depends a lot on what you intend to do with the product and how expensive it is to bring to market. In the Alexander Graham Bell case, consider how expensive it would have been to set up all that infrastructure before you could charge dollar one for any service. Even for the example Farmandfamily gave, I'm sure the $30k+ spent by the invention thief was mostly to make the product and bring it to market. A patent is a small portion of that. For the windshield wiper inventor, if he had NOT had a patent, I very much doubt the car companies would have even made that initial offer of a few million dollars.
The set of circumstances where it would make sense for you to get a patent is a fairly narrow, but not nonexistant one. You don't necessarily have to sue to get some value out of a patent, if your most likely competitors are small, domestic ones simply sending a threatening letter saying you have a patent and they are infringing may shut them down. If your most likely competitors are, like the case with the Yeti coolers, large companies that can get it made cheaper in China, you will need to be partnered with a large company to have a chance of fighting them off. Clearly Yeti is either not large enough or doesn't have patents broad enough to protect its designs.
The best option for an individual inventor is probably, as the Living Stingy article alludes to, partnering with a larger company with the resources to 1. Make, 2. Market, and 3. Sell your invention, and take a few percent licensing fee in return. They may have more interest if you have patent protection, as they could be a more credible threat to an infringer than you as an individual. They are also unlikely to try to end-run you if you have a patent filed, because it is probably cheaper to pay you a few % than to take a risk that you get mad and sue them, juries actually tend to be pretty sympathetic to damage cases like heartless corporations stealing from garage inventors.
That said, most companies that actually develop new products usually have more products they would like to develop than they have money for, so it can be hard to get them interested in an outside idea. Having at least a provisional on file, plus a NDA, would be wise before showing them anything. Then document every interaction.
Do be careful with showing this to anyone not under NDA, or offering it for sale, or putting it on a web site to show it off. Any of those things starts a 1 year clock in the U.S., if you don't file at least a provisional during that time, you can't ever get a patent on it. Some of those actions immediately bar getting a patent in some foreign countries.
If you do decide to pursue a provisional, which is low cost, you want to try to get your product manufacturing and potential partner companies lined up pretty quickly, so you, or they, can decide on the more expensive follow-on utility filing during that 1 year window. I suggest understanding how to do a good provisional application, or its largely worthless.
I typically have a good half hour conversation with anyone thinking of getting a patent from me so they understand the pros and cons. I would rather not waste my time and their money if it doesn't make sense for them.
