How to estimate future App Store sales?

Before launching an app in Apple's iPhone App Store, it would be useful to have at least a vague idea of the sales potential. But how on earth am I going to estimate this?

As an example, let's take a small productivity app with a certain detail that makes it unique and desirable, at least to your niche market that is productivity-oriented iDevice owners.
You could sell this app for the obligatory $1.
For such a product, you would have a pretty clear idea of how to sell it; elevator pitch, typical user profile, etc.
But we have no idea about the customer demographics in the App Store.

This could sell 20 copies in a year, or 20 000, or even 20 million -- who knows? Given that there are more than 300 million iDevices around and 35 million activated last quarter, the market potential is staggering. Even getting literally one customer in a million would bring income.

How can I get an idea of what to expect, in terms of copies sold?

I have no idea how to estimate the amount of sales you are going to get, but I suspect with so many variables there just isn't a way to have a number with any relevant level of accuracy.

If you just want a number to make you feel better than choose a number that does that. If you want to an actual number that's realistic so you can make decisions based off it then I think you are going to be disappointed.

Do you really need this? If it's for pricing, I wouldn't base it off this number anyway. You can't just price based on how much money you would like to make. I would price it based on other factors. I can't really think of a good reason to need this number, just launch it and find out?

If you're not able to estimate this for your market for your product, then

1. People outside of that market and that product won't be able to, and
2. You need to get some understanding of the basics of product lifecycle.

There are fairly basic processes around product lifecycle, evaluating ideas, protoptyping, gathering feedback, evaluating markets, performing market research and generally evolving in stages using iterations of each parts to improve it and continually analyse its viability as a commercial product.

The idea of "wait and see" in business is, frankly, absurd, unless the time you spend developing the product is worth nothing.

This question is really "what is the breakdown of the entire process of product development" and I suspect that you have done some parts of this process implicitly (like you are solving a specific problem in a domain you have expertise in). I'd just suggest being a little more formal in the other parts of the marketing process, or get someone with more experience in that area involved.