I've historically had a number of different approaches to how I plan my novels. For the record, I've begun something like 24 different novels that never panned out. 19 of those came before I finished my first novel, CAYENNE, in 2001. I got as many as 45,000 words into some of those books, but for some reason all of them petered out until I got to CAYENNE.
What changed? Planning. With CAYENNE, for the first time in my career I did detailed planning and outlining before I began the novel. I had about 25 single-spaced pages of detailed scene plans and such for that book, and during the year and a half it took me to write the story, I didn't deviate much from that central plan. That turned out to ultimately have a negative impact on the believability of the story (due to faults and oversights in my original plan), but it did aid me significantly in writing the book. CAYENNE turned out not to be worth publishing, but writing it was a great step in the right direction for me. It helped me learn some of the patterns of being a real novelist.
I was generally preoccupied with my career as a software programmer from 2001-2004, but I did make a about six halfhearted attempts at writing a second novel. Three of them I planned extensively, but those plans didn't resonate well enough with me, and so I ultimately dropped those stories. The other three I did not plan at all beyond the basic premise, and they died even faster.
It was in mid 2004 that I had some of the first ideas for THE GUARDIAN, and I spent a good number of months jotting down ideas and plans for a series based around those concepts. I didn't write any actual narrative at all, or any particular low-level plot, but I did write some 40 pages of world planning, magic design (it's a fantasy), character development--all that backstory stuff.
In early 2005, I actually started working on the narrative of the book itself. At that point, I had loads and loads of planning done, and I knew many of the places, characters, devices, etc, extremely well. I also knew the shape of the over-arching story and themes. What did I not have? Actual play-by-play notes for the individual chapters. I struggled through about 20,000 words in this manner, but the going was very slow. I was basically planning as I went, and that worked pretty badly. I mean, the result was fine, but it was agonizing for me and quite slow.
But, because of all the other planning I had done and my commitment to the story, I didn't let that stop me. Instead I shifted tack, and started planning out just the next chapter or two before I wrote them. This was a little bit awkward at first, but over the course of the rest of the book I got very used to this approach. For me, it turned out to be the perfect hybrid--I didn't have to try to mix planning and writing together as one activity (agony and slowness), and I didn't have to plan out the entire book in advance.
The top advantage of not having the entire book planned out at the start is that it gives you flexibility to do the unexpected. If you've planned out your component parts (characters, settings, overall conflicts and motivations, world rules, etc.) well enough, then this sort of approach gives you room to just let the characters be themselves and take you somewhere unexpected. Trust me--I would never have planned on the things that happen in the second half of THE GUARDIAN. There would not have been a character named Valdur if I hadn't taken this approach with my outline. And it's exactly that character, and the events surrounding him in the second half of the book, that readers find so compelling. It would have been an okay book with my original general ideas for the book, but nothing all that special.
Truth be told, a lot of the stuff I originally planned just wasn't that interesting. It set up a nice framework, but the really unique stuff developed on its own because I gave it the time and space to do so. But on the other hand, I never would have made it to the latter half of the book without all the preplanning I did. For me, it takes a little of both approaches.
I recently saw an interesting post over at Caitlín R. Kiernan's blog that was comparing her original proposal for DAUGHTER OF HOUNDS to what the actual finished product was. I confess to not having read the book, but it sounds as if Kiernan is pretty glad that the finished product was divergent. It sounds like she experienced the same sort of "transformation on the page" that I did.
I've read what a lot of writers think on the subject of outlines, and everyone thinks something different. Like anything else with writing, I think the best answer is that you have to find out what works for you personally. There aren't any magic formulas. But one other thing has also become clear to me: it seems that nearly every writer uses some form of outline, be it the full deal, a matrix, or just a starting synopsis. If you're trying to decide on some form of outline, I think the best way to approach it is to consider what form of outline might have the best payoff of usefulness, while having the lowest cost in terms of distraction from the actual business of writing.
Because that's what it's all about. Letting yourself write. The reason I need the kind of outlining I do is simply that not having it distracts me while I'm trying to write. It's hard to even make myself sit down to type if I don't have some sort of plan. For me, that makes the outline absolutely invaluable. But to each his (or her) own.