As a current PhD applicant who's spoken with a number of faculty, doctoral students, and recent graduates, I'm going to have to disagree with AleksandrM.
You have to show a wide knowledge of the area you would like to study, as well as a specialized understanding of the relationships between what you want to focus on and the value of such research.
That could not be more incorrect. The faculty certainly do not expect you to have a wide base of knowledge BEFORE you get your PhD, nor do they expect you to have "a specialized understanding" of the intricacies of research topics. If you had all that, why would you even need a PhD, and why in the heck would it take 4-5 years to complete?
The most important thing is NOT having a well-defined research proposal. As a matter of fact, having a research idea that is too well-defined can very well work against you. Not only do most people tend to study something completely different than they thought they would when they entered the program, but being too specific also exposes you to danger on a number of levels.
First, you run the risk of showing your ignorance. Let's be honest -- most applicants are not going to be able to have a high quality research proposal complete with methodology formulated at the time they apply. Not only are they unfamiliar with the breadth and depth of the literature, they are also not yet trained to be researchers! Coming to the plate with an idea and presenting it as certainty makes you look too cocky and unaware of your current ignorance.
Second, you run the risk of being a poor match! If there is no faculty member there who is interested in your extremely specific research proposal, nobody will want to work with you.
Your best bet is to mention specific topics that interest you and why. For example, you're interested in consumer choice models because you've always wondered how much impact SKU proliferation has on consumer purchase behavior and you like Bayesian statistics; or you're interested in the role of self-image as it relates to consumer perception of brand authenticity because of something you learned in a psychology class. Then, if you want to take a little bit of a risk, mention specific faculty at the program and their research that broadly relates to your interests.
In the end, your SOP and LORs have the biggest impact on your candidacy. GMAT/GRE scores largely serve as screening mechanisms up front and as differentiators again later in the process when choosing among short-listed candidates.
Your best bet to gain admission to a top PhD program is to score in the 720+ range, have a 3.50+ undergrad GPA and/or VERY high graduate GPA (3.8+), very strong LORs from well-known faculty, and a very well-written SOP that clearly addresses why you're interested in doing research and why you're well-suited for a career in academia.
If you, for any reason, mention that you are primarily interested in teaching or that you are interested in returning to work in industry, you will not be admitted anywhere, even if you have an 800 GMAT and a 4.0 from Harvard.
From one aspiring marketing PhD to another, best of luck!
