By this time, you should have a test site incorporating everything you've learned up to now. It doesn't need to be complete, but it does need to have enough to give you useful information about how your users are going to interact with your site. So you're probably talking about a prototype that's at least 2-3 pages of content, which includes a way for a user to accomplish something they've come to your site to do. Even if the functionality isn't there (for example, if it's a store without a shopping cart), it can still tell you plenty about how successfully your project is going. Just remember that you should still user-test any missing functionality later on before the big unveiling.
Knowing what you do about your users, ask them for feedback of your prototype. If you already have a mailing list, that's perfect. If not, use your earlier research about your likely users to find them and reach out for their help.
You could ask questions along these lines:
- Was the site easy to navigate? Were you ever lost?
- Was any part of the site confusing?
- Do you like the look and feel of the site?
- If it were up to you, what would you change?
- Do you understand what the site is for and what you can do there?
You should also include a few questions that willl confirm for you that the person answering the questions is part of your target market (age, gender, basic interests). This will help you understand the data better, and it has the added benefit of possibly uncovering a market you didn't know you had.
You should continue to gather data from user tests until you can be reasonably confident that you are on the right track with your design and with what the site is offering its users. If at first you aren't getting the results you desire, you can pause testing, revise the site, and begin a new testing round. Repeat until the numbers are looking good and the revelations are few!