During the development of an idea, it's easy to lose sight of your original vision. You begin to see new things in your work, and it carries you in new directions. Often during this process you end up with something very different from that which you originally intended to create. During the process, you forget the problem that you set out to solve, and judge your work purely on its aesthetics, or the feeling of the moment.
The creative brief is the solution and recognition of this situation.
A creative brief is a written, detailed description of the design problem to be tackled. It is not simply a statement of the problem (a brochure needs to be created), but also gives you relevant facts about the client, their target market, the market goals they wish to achieve, their history as a company, the intended message which needs to be conveyed, and any other relevant information. A good creative breif gets beyond the simple question of "What do you want", and instead sets in a number of parameters which define a "space", what is often referred to as the "Problem Space". Within this defined space exists viable solutions, and beyond it lay non solutions. In theory. A really good design process will recognize when a creative breif is inadequate, and midway through redefine the problem, perhaps in different words, perhaps from a different perspective. For instance, a creative brief might stipulate that the client needs a business card for her florist business. A designer, after some meetings with the client, might redefine the problem as "The client wishes to attract more business from engaged couples, and stay within a budget of under $4000"
The key here, is the more detailed you get, the more you crystalize a vision of what the problem *really* is, the more productive the design process becomes.
Through each revision, and at the end of the process, ask yourself: "Does this *really* solve the problem?". be thoughtful and honest to yourself (or others) in giving the answer. If you answer no, ask yourself why, and you will understand what revisions need to be done. If you answer yes, be careful that you are not deceiving yourself. There is no such thing as a perfect solution- But within a problem space there are many possible solutions. Is this the best one?
This may seem stupidly obvious, but in this complex mesh of theories and processes that is the design process, it is easy to lose sight of what is obvious, and get lost in the details of implementation.