These are very similar questions. My freelancing routine isn’t much of a routine. As a general rule, though, on a week night, after dinner I settle into my messy desk with a big glass of water and maybe some scotch, fire up the laptop, and look at what needs to be done. Then I write.
Process-wise, I start in a notebook where I usually write things by hand. This is more a convenience feature than a preference– I always have a notebook with me, and it’s very fast for me to write on it, compared to how slow it is for me to fight with Autocorrect on my phone. (Seriously, the lack of a keyboard on phones is the #1 thing keeping from using it as my main computer.)
From the notebook, I start an outline in MS Word or Google docs, depending on where I intend to share the work. Usually I go with Word, because it’s extremely familiar to me.
I work from an outline. I’m meticulous about outlining and using Word’s comments and tracking features to annotate what needs to be done or what needs to change in a document. That said, I still miss a lot of little things when I’m working.
From there, I do three or four readthroughs, particularly of things that players will see during a playtest. I add a version number to the footer, as well as the date. Print off a copy and take it to its first playtest, where I use the same notebook as before to make notes about feedback and write down my playtesters’ names. When that playtest is over, I take my notes back to the Word document and transfer them into the document in some way that’s meaningful. From this point on, every round of playtesting and revision increments the version number by .1, until there’s a major rules revision. Threadbare was version 3.5 when it finally went to press, and the post-press errata technically makes it 3.6.
There’s another round of writing and outlining and more writing. Usually there’s another round of playtesting as well. Then it goes to the beta playtester circle on Google+ for people to read and playtest and send their feedback.
What follows is more playtesting, unless there’s a deadline. For a new game, I try to take it to at least one convention for playtesting, and will usually run at least one online game to reach a wider audience as well.
At some point, it’s time to look into releasing the game. I either lay some money down for artwork, submit the game to a publisher or contest, or post it without artwork to the blog.