2. Choose the wrong architecture for the book. This is important because, while you think you're going to Mars, you'll actually be heading for Uranus.
3. Do lots of time-consuming primary research that you will throw away later.
4. Stop a year into the research process to write another book about, say, the history of modern air conditioning.
5. Realize that some of what you learned writing the other book could be used for the book you were already writing, but in the middle section. So, research and write the middle section of your book. And, as you put the cart before the horse, carefully maintain the wrong architecture.