I certainly like the idea of not offering a free plan at all and it's definitely the way I'm intending to go (at least initially). One of the advantages I can see is that it positions your service as a serious app targeted at professionals (which, in a way, encourages professionals to use it). Also, even if free users cost you nothing up front, you still have to store their data and support them. I'd rather 10 paying customers than 1000 free + 10 paying.
As you mentioned, if free users contribute (perhaps, depending on the context, by making their data public) then a free plan becomes more useful.
My own advice would be to have only one paid plan. Some sites have multiple plans at multiple prices and try to segment their user market. It will save you a lot of time hassle in development if you just have one premium plan to start. When you are a success you can look at user segmentation.
Also - be aware that if there is a community element to your service multiple plans can cause discord.
1) Get good at SEO. SEOmoz.org is the best place to start. 2) First, prove that you can bring in 100 people and have some meaningful percentage turn into customers. Might take a lot of design/wording iteration-- might not. Depends on the product simplicity/value prop. 3) Price it in such a way that you can afford marketing. If you can convert 3-5% of your visitors into customers and you have a reasonable high price point, you can afford adwords, which is a huge win. 4) Understand the "lifetime value of a customer". What's the average customer worth? How long do they stay with you? This will tell you how much you can spend per user on marketing efforts.
Let's see... Other generic advice:
1) Don't offer a free plan unless free users provide some other value. 2) DO offer a free trial. It seems standard practice to ask for the credit card # at the time of the trial... If your service is sticky enough, you might consider doing it the other way. 3) Use BrainTree payment services. They've treated us pretty well. 4) Price it high, but offer a "seasonal discount" so that people feel like they have a bargain if they strike now. 5) Measure like crazy. Understand your funnel REALLY well.