How can I measure my future audience?

Before taking on a project, there is always a question: "How many people will use this?" So how do I measure that "people", at least inaccurate.

For something quick, dirty, and fairly cheap, try something like the following process:

1. Research potential competitors and their products and/or services and realistically assess whether or not you can provide a unique selling proposition that differs enough from the competition.
2. Use Google Trends, the
   External Keyword Tool, and
   the Traffic Estimator to get an
   idea of how popular your project,
   potential competitors, and the
   general industry might be. If it
   seems popular enough, move on to
   the next steps.
3. Create an introduction video to your
   product idea describing it and
   perhaps some concept art. You can do this for free even without a video camera if you have a microphone and PowerPoint. Just do a screencast using CamStudio, an open source screencast software.
4. Sign up to an email newsletter
   service like MailChimp or Campaign
   Monitor. Both have free signup and either no or very minimal cost below a certain number of subscribers/email sent.
5. Create a landing page on a website with the video, the newsletter signup form, and perhaps a written description as well. At the end of the video, indicate that if they want to be informed when the product comes out, they should sign up (for free) to the newsletter.
6. Create a Google AdWords campaign that directs people to your landing page. The ad should read the same as if the product actually existed. Depending on the industry, you can probably get a statistically relevant sample for a few hundred dollars or less.
7. Measure the level of interest based on the number of views of the ad, the number of click-throughs, the bounce rate, and the number of people who sign up during a given time period or ad budget. If the numbers are promising, then you can move on.
8. Create a Minimum Viable Product and begin selling it. If sales go well, begin adding features and developing it further.

@vituosi makes some great points, but I use an easier approach. Take your most realistic competitor, (similar product, pricing). Choose a company that has not been around 10 years! I usually choose the poorest performing company with the same pricing. Not a hobby, but a real company struggling to make it.

Then find out how many customers they have. Some of them will actually tell you if you call them and pretend you want to sign up, write an article, or you can do some digging (for example web hosts by how many sites on the IP block)

Last, take the number of clients, multiply it times their lowest price point, and then divide it by two!

That low number will give you your WORST CASE scenario. I always use a worst case to measure if a product is viable, and also to keep development within budget and time. This helps you look close at the comeptitors, and use them a gauge of your success.