We have established that we aren’t our users and that users can be unpredictable in how they use our site. So the best option is to find out how users behave with our site as early on as possible. Therefore, we should launch with the Minimum Viable Product (MVP) in order to get it in front of users to determine its future.
What the MVP is varies by product or service. If you are selling software to businesses, then perhaps its just paper prototypes: photoshop comps depicting what the product will look like. You can show them to potential customers to see if they are interested as well as get feedback from them before you even write a line of code. If you are creating a new website to soothe a pain point, then create just that functionality and don’t worry about the other standard functionality that sites in your market need to have. If you are accurately soothing the pain, then users will respond to your offering; if you aren’t, then they won’t.
Don’t spend months or years developing a product, scared to bring it to market and have it fail; scared that once it’s visible some other company will copy it and steal your potential customers. Get the product out there as soon as possible to test its validity. You need to verify that there is a market for your product and you don’t need to build out every piece of functionality to do so.
Update: Some great examples of Minimum Viable Products in action