1. The most important thing, of course, is choosing the topic that's right for you (and that your thesis director doesn't absolutely hate). There is always a temptation to choose a subject that is fashionable right now, will look good to a hiring committee, or will be easier to turn into a book. You need to remember, however, that before the hiring committee or the publishers even get to hear about it, you will have to spend months of your life living and breathing this topic. If it doesn't inspire you a lot before you begin wiritng, there is a danger that your enthusiasm for it will peter out pretty soon.
2. Decide how many months you give yourself to do pure research before you start writing. Remember, you most definitely don't need to read every source available on your topic before you begin to write. When you start writing, you will find out that the process of writing will always lead you to look for more specific sources.
3. Every time you read what might become a primary or a secondary source, I suggest that you copy the quotes that seem like they could be useful onto a cue card. Don't leave it for sometime later on, or you will find yourself in this unpleasant situation (known to every researcher) where you have a nagging feeling that you read something somewhere about something important but cannot remember where it was. I always carried a stack of cue cards with quotes on a key chain and went through them regularly. Eventually, I arrived at a point where I knew my sources so well that I didn't need to stop and rummage around in articles and books every time I wanted to quote something. This allows for a much faster and more seamless writing pace.4. One of my professors once told me that the best pace to write is 3 pages a day 5 days a week. I found it to be a great suggestion. This writing schedule leaves you with 10 good pages a week (after revisions) and plenty of time to explore more primary and secondary sources. I know there is always a temptation to produce 8 pages in a row on a good day when you feel inspired. In my experience, however, this feat would leave me pretty useless the next day and those 8 pages would turn out to be pretty bad anyways. Besides, you always need some distance from your writing, so there should be at least one day a week when you don't work on the dissertation at all. Otherwise, you might start hating it pretty soon.
5. Learn to let go. As hard as it is to accept it, there will always be more relevant sources, critical theories, and primary texts that can be added to the dissertation. So at some point it's ok to just leave it the way it is. There will always be time to add more when you are transforming it into a book or a set of articles.