I've done plenty of writing for Paizo's Player Companion line: Magical Marketplace, Champions of Balance, Giant Hunter's Handbook, Melee Tactics Toolbox, Black Markets, and the soon-to-be-released Blood of Shadows. Although there is some world-specific flavor in these books, they're primarily aimed at giving new rules for players: new spells, feats, traits, equipment, magic items, and sometimes entirely new rules subsystems. They're very crunch-heavy.

In each case, the process (for me) goes a bit like this.

The line developer (previously Patrick Renie, now Owen K.C. Stephens, both great guys who juggle a LOT of work and have really keen insights into the game) send out an outline that talks about what the product is going to be, and all the sections they need (often, a two-page spread), along with their word count. Sometimes, there's a lot of material in the outlines already about the sections. A hypothetical book about oozes would identify a two-page "Ooze Spells" section, a two-page "Famous Oozes of Golarion" section, an "Ooze Equipment" section, and so forth, and might have five or six paragraphs under each:  "Ooze Equipment" might say a lot about how what types of equipment should go in here ("not just oozy items, but items to control, trap, and train oozes") and even some specific suggestions ("include some sort of saddle that lets you ride oozes, and is resistant to the acid damage that some oozes can do; include a netting of some material that, when used to grapple an ooze, keeps it from using its split ability").

I'll look all these sections over, decide what seems the most interesting and how many words I think I can produce in the allotted time, and respond with my preferences. The developer gets preferences from all the authors (sometimes as many as 4 or 5 people) and then re-sends a formal outline with everything assigned. Often, he's able to get me what I asked for, but not always.

For Black Markets, I got the sections about the Corpse Trade (a discussion on the use of undead as servants and a new type of magic item called necrografts), Secret Faith (magic items usable by characters who want to keep their faith hidden) , and Black Market Magic (new spells). i don't remember what I asked for, but those three were all awesome and seemed fun.

Then I start brainstorming. In my much-abused leather-covered notebook, I'll jot down ideas for each of these. I'll be sure to spend at least a few days brainstorming, coming up with lots of ideas and refining and revising the ideas I have. (Boring meetings at work are a good place to do this.) When I think I've got a ton of ideas--way more than I'll need--I start researching.

Researching is critical, because it lets me know if my ideas have already been covered somewhere. Although in my writing I usually stick to the official rules source of paizo.com/prd, that website doesn't contain materials that have been in the very many softcover books Paizo has produced (including other Player Companion products). So I use d20pfsrd and the Archives of Nethys too. This research lets me know if I need to jettison any ideas that have been wholly traveled before, or whether I need to refine any of my ideas to distinguish them from rules in existence. This part is often hard, and not my favorite part, but it's really important. When I miss something, I'm sure to hear about it, and feel bad (such as my secret coffer spell in Black Markets, which was too similar to a spell that already existed!).

This is also where I keep a careful eye on how powers are worded. If I create a magic item that is similar to an inquisitor's class ability, for example, I note how the class ability is worded, and try to stick close to that. The more similarities within a large rules system, the easier they'll be to understand.

Often, the outline asks that I submit my brainstorm ideas for approval before writing them up. I had what I thought was a great spell idea called dogsbody, which summons a small automaton, like an urchin, to perform simple services the same way that mount summons a useful steed. My developer thought it would need so much clarification and such that it would be too complicated. So I didn't proceed on that.

Then, I dig in to the writing, putting together all of the items I've brainstormed, in this rough order: all the specific items suggested in the outline, the items I brainstormed that I thought were pretty awesome, the items I brainstormed that didn't seem quite so awesome but add something interesting to the corpus of rules. In an ideal world, that's where I'd stop, but that usually falls short of my word count (it didn't in my last Player Companion project, but that was an anomaly). So I have to dig back into my brainstorming notes and convince myself that something I'd jettisoned for being silly, useless, or duplicative of something else might be useful. So I revise those things and inch a little closer to my word count. I usually have to do this a few times, and include a few things I'm not completely proud of, but they get the job done.

(At some point in all this, I get the contract for the word count allocated to me, and I send it back in. That's the administrative side.)

Once I've got everything pulled together in approximately the right word count, I go over it all and revise for corrections, look up lingering rules items (can you say "a curse aura" rather than "a curse upon an object or creature"? Are feats written in the first person or third person?), and generally trying to inject more awesomeness. As an example, my necrografts section of Black Markets was pretty solid, but upon review I thought of tying each type of necrograft to a specific type of undead: mohrg tongue, wight hide, and so on. More awesome.

If I can, I bounce some ideas off of friends of mine or other authors I'm working on the project with, to make sure I have power levels right or referencing the right things (such as, "this is a neat magic item, but it should have the know the enemy spell as a prerequisite rather than identify").

Then I give the whole thing to my wife for her close read and edits, usually on the Saturday or Sunday before it's due (most projects are due on a Monday, and I prefer to send them in on Sunday evening or before Paizo's offices open on Monday morning--easy to do because I'm two hours ahead of them). Because I've already been over the entire thing at least once, her edits are usually minor, but she invariably finds words I can trim or punctuation I've messed up. I make her edits, submit, and I'm done!

The final part--for me--is when the book is released. It's exciting to me to see what's changed; often it's not much, but I keep in mind any revisions Paizo had to make in order to improve my writing in the future. More often, it's something that's been cut for space or moved into another section. As an example, I had a 0-level spell called vacuous vessel in Black Markets that I was sad to see cut from my Black Market Magic section, only to notice it included in a different section (that I hadn't written) about smuggling. So it was still there!

I also lurk on the product pages on Paizo to see if people raise questions or submit reviews that touch on the work I specifically contributed, and answer questions when I can.

The Player Companion writing is a lot of work; probably more work per word than any other writing I do. But it's a great way to keep flexing my research skills and design skills, so I hope to keep doing more of it!