I often hear the touting of project-based billing, but the reason I'm afraid to make the jump is that in my field (developing prototypes, often novel, for scien

I often hear the touting of project-based billing, but the reason I'm afraid to make the jump is that in my field (developing prototypes, often novel, for scientific applications), the requirements are just not well-defined enough to give an accurate estimate. If I went project-based, just to cover the risks involved, I might have to charge 2x or 3x what it might actually take me. This may work if your clients are big corporations for whom 10k is the same as 20k is the same as 30k, but not when you're working with startups, boot-strappers, etc. I know the next thing people will say is find better clients, but some people just have expertise in areas where the clients tend to be smaller and more cash-strapped. Thankfully the work is interesting. Right now I give estimates, but still provide hourly billing as a safety net. Once I'm confident that my estimates are consistently accurate, I might move to project based billing, but it also feels a bit sleazy because what I would do at that point is just take ESTIMATE x SOME MULTIPLIER. In other words, project-based billing sometimes sounds like a hand-wavy way to charge the customer more without them being able to tell.