People assume books get written in long uninterrupted stretches, ideally somewhere with a view. Mine get written between a class that ended at 11:20 and a meeting that starts at one. I've published textbooks with Pearson while carrying a full teaching load and a coordinator's calendar, and the only reason it worked is that I stopped waiting for the long stretch.
The unit that works for me is about ninety minutes. Less than an hour and I spend the whole session remembering where I was. More than two hours and I start editing what I've just written instead of writing what comes next, which feels productive and isn't.
The real cost isn't the writing time, it's the reloading. A chapter is a structure you hold in your head, and every time you set it down you have to pick it back up. If four days pass, the pickup takes most of the session. If it's a day, the pickup takes ten minutes. So the frequency matters more than the length. Three ninety-minute sessions in one week beats a full Saturday, and it isn't close.
What makes reloading cheaper is leaving yourself a note. I stop mid-section rather than at a clean break, and I write two lines at the bottom about what I was about to say. It feels wrong to leave a paragraph unfinished. It's the single most useful habit I have, because the next session starts with an instruction instead of a blank page.
Textbook writing has a structure that helps here. The chapter outline exists before the prose does, and it's been reviewed by people who will tell you if it's wrong. So on any given afternoon I'm not deciding what the book is about, I'm writing a section I already agreed to write. That removes the decision that would otherwise eat the ninety minutes.
The memoir doesn't work that way, and the difference has been instructive. There's no outline anyone approved, no reviewer waiting, no deadline that comes with a contract attached. The same ninety minutes produce less, and some days produce nothing I keep. I've had to build the scaffolding myself, which mostly means deciding the night before what I'm going to work on, so the morning doesn't get spent choosing.
The other thing I've had to accept is that teaching terms have a shape and the writing has to fit inside it. September and January are unusable. Reading week is worth more than the whole month around it. The weeks after grades are submitted are the best writing weeks of the year and I now plan for them deliberately rather than discovering them by accident.
None of this is a productivity system, and I'm wary of turning it into one. It's closer to a set of concessions. I concede that I won't get the long stretch, that the mood won't arrive on schedule, and that a book written in pieces will need more revision at the end to make the seams disappear. That last part is real. Work produced ninety minutes at a time has a tendency to restate itself, because you forget what you already said. I spend a lot of the final pass deleting the second explanation of things.
What I'd say to anyone trying to write around a job is that the schedule matters more than the inspiration, and the note you leave yourself matters more than either. The book gets written in the gaps or it doesn't get written. There aren't going to be better conditions later.