I don't mind answering tangential questions of this kind, so that's fine.
First, most of my books aren't strictly speaking academic: they are on academic subjects, but for general readers. There are only two full-length things I've written which would be aimed at an academic readership: one is an encyclopaedia, and the other is my PhD thesis, which I haven't had published yet (although some bits have been turned into articles). I'm also going to be editing a volume of papers for an academic press, so that will be properly academic, but editing such a book is very different from writing one. I think that writing something like an academic monograph will necessarily be somewhat different from writing (say) an introduction to or overview of a subject, simply because of the different kind of research you'd need to do - as my answer to the previous question should make clear.
But that said, I think that the skills involved in writing are pretty much the same no matter what you're writing (sticking within a non-fiction, vaguely academic field, at any rate). And you're right to identity organisation as the big problem that everyone faces. It's all very well have lots of stuff to write about, but what order does it go in? The first thing to realise is that there is almost never a single right answer to this. There are reasons compelling you to organise it this way, and reasons for organising it the other way, and you have to just try to judge which way is most successful.
As an illustration, the last book I did was on the history of the early church, and the final chapters described (among other things) the Arian controversy and the increasing measures taken by the state against paganism. Now these are very different subjects, and it made sense to treat them separately, even in different chapters. But at the same time, they were historically very closely interwoven: for example, Julian's decree that exiled bishops could return home played a key role in the development of the Arian controversy, but it occurred as part of his strategy to favour paganism over Christianity. To keep the two subjects separate seemed to require describing the imperial succession twice, since emperors meddled in both areas. But to mix them would have been to make a very complicated narrative. In the end I had to compromise. The first part of the Arian controversy is told by itself, and the second part is interspersed with the account of the paganism issue. There is still some repetition but it is acceptable. This kind of thing is very difficult and sometimes you just have to shuffle material around until it reads relatively smoothly, and that of course takes a lot of time as well. In fact, rewriting takes longer than writing. And I, at least, spend a lot more time cutting words out than I do putting them in in the first place. This is a truly horrendous activity but it has to be done.
However, you very much do have to have a plan before you start. This is true not only from the point of view of writing the project but also from the even more pragmatic point of view of publishing it. I wouldn't recommend writing a book of this kind until you have a contract to publish it (unless you're just doing it purely for fun without an eye to publication), because any publisher who is interested in it will have a particular way of doing it that they want to do. What I mean is, a publisher may wish to produce a book on a certain subject, pitched at a certain readership, of a certain length, covering certain topics. If that is similar to what you're interested in, then you're in luck. You can write a proposal for such a book without great difficulty, and then if the publisher agrees, you can set about writing it. But if you write a book first and then pitch it about, the chances of finding a publisher who want to publish precisely that book are very small.
The way publishing works is this: it all begins with the proposal. Either you write a proposal for something you want to do and send it to every publisher who might be interested, or, if you're lucky and you are already known to a publisher, you might meet with them and discuss what projects you might both be interested in, and they ask you to write a proposal. Either way, you have to write a proposal. That proposal must include information such as a proposed list of chapters, word length, prospective audience, description of already existing rival titles, and so on. It may also include a sample chapter. Then the publisher agrees to it, you sign a contract, and it's time to start writing. Or they don't agree to it and you must refine it.
The way I got into publishing was that I had an idea for a book, so I wrote a proposal and some sample material and sent it to every publisher I could think of. I then got a vast number of rejections - from those that bothered to reply at all - but one publisher was interested in a quite different sort of book but thought I might be good at doing it. So I wrote a proposal for that, and they requested changes, and this went back and forth for quite a while until finally I got rather sick of it (they were not the most efficient publisher) and I took the revised proposal and sent that all round the houses again. This time two publishers liked it and I ended up with two contracts, because they wanted sufficiently different variations on the original idea. Only after all of this (which took about eighteen months altogether) did I actually begin writing the book. The first publisher liked what I did, and all the books I've done for them since have been ones that they initially proposed to me, although I still had to write proposals and get them approved. With the other publisher, I've only just finished that book - it took eight years! That was rather extreme, but writing is a very, very time-consuming business (I've spent the last couple of weeks merely compiling the index to the eight-year book, and that's not a task for the faint of heart). My main publisher recently proposed a new idea to me but I had to turn it down this time because I just can't guarantee the time to work on it: if I have that much time, I should be devoting it to articles.
Having a proposal is necessary not only to get a publishing contract but also to have a good sense of how the book is to be set out and organised. With some projects, most of the organisational work is done at the proposal stage rather than the actual writing - indeed the writing is almost fairly mechanical, just a matter of filling in the material that the proposal anticipates. Of course you always find that changes are required, but the more work you put into the proposal stage the less hassle it is organising the actual writing. I still always write far too much and end up having to cut lots of it out though - I'm not very good at sticking to work limits on the first draft.
As for noticing stuff that might be useful in future projects, I don't have any system beyond just remembering it.
If you're keen to write then I definitely encourage you to do so, and it's very important to keep practising because writing really does improve with practice. Mine has, at least. It does also force you to develop good habits of organisation and motivation which, if you do a PhD, will stand you in very good stead there. As I say, what you should do first is develop a good proposal and send it to all the publishers who might be interested in it. Make sure that they all do publish that kind of book, though. The biggest mistake that people make when trying to publish things (and this applies to academic articles too) is to send them to publishers who simply don't publish that kind of thing.