First step: Need to create something. I don't generally like what comes of creating for the sake of creating (which is, I think, responsible for much of the junk that's on the market)
As already said, I generally either have a routine I'd like to develop, a scene I want to create, or a premise I want to explore.
Sometimes they all come together, but one of those is usually the motivation.
I will often create a script at that point. Before I decide on any methods, I decide how I want it to look to the audience.
Often at this point I will tell my wife about it. If she thinks it makes sense and sounds interesting, I continue. If it doesn't make sense, I rewrite it until it does. If it doesn't sound interesting, I tweak it until it does, or if I think it's interesting enough anyway, I just continue.
Then I brainstorm methods to use to create that scene/image from the audience's perspective. I figure out a method that will work, then I do several rounds of playing it through in my imagination to see if I can spot potential problems. After I've made the adjustments I need to make, I re-write the script and blocking to accommodate those changes.
Then I put it in front of real people.
Then I go back and change what didn't work and keep tweaking until it's something that I enjoy performing.
The process of creating a method is about taking what you know already and sometimes applying some lateral thinking to achieve the goal. Also, don't be afraid to change ... well, everything. Method, premise, presentation, script, blocking ... The first run at a new routine is almost never what it will look like when I'm performing it regularly.