This was my process. The outline is very long but fairly simple and hopefully it will at least give you some ideas. It also presumes you will not have an editor or friend that will give real critical feedback for at least 6 months.
——
Write your first complete story, ideally one that is under 2000 words. Something basic, short, and familiar without being a ripoff of anything in particular.
Then keep writing stories that way until you’ve forgotten most of the details of the first story. read it now from an outsider perspective and note the things that don’t work.
The most common offenders early on will almost certainly be phrases or modifiers that are repeated too often(“to be quite honest”, “suddenly, x happened”) as well as sentence structures that do not flow well or require re-reading to parse. This is all easy to fix; simplify the complex sentences and substitute common phrases to expand your prose.
After quick and dirty adjustments, keep writing new short stories. Then reread again and adjust, again. Eventually you’ll have a “library” of like 10-20 stories that you’ll either know way too well to acquire the easy outsider perspective, or have hammered into an acceptable quality.
At this point writing ~specific~ short stories is a good idea. For me, I focused on methods. So it looked like this:
-Start with a strong visual and go
-Start with an ending and go
-Start with a strong emotional event and go
-Start with one well defined character and go
-Start with a writing style/intention, mimic a real author
-Start with an intended audience reaction
generally the later attempts should be more difficult and more specific. For example, write the same story twice from radically different perspectives. Force out a story in iambic pentameter. Things like that.
Batch 2 does 2 things fairly organically. One, it implicitly teaches you what KIND of writing you enjoy while helping you hone your own voice. Secondly, it forces you to examine the actual structure and components of a story and how configurable they are without diving into anything formal or educational. That last part was important to me because it’s VERY easy to absorb too much of an authors style by listening to them talk about writing, and formal education takes all the fun out of writing.
Worth noting; this second batch of stories will probably suck. hard. Worse than the first. They are handicapped and probably very difficult to complete well. That’s okay. Examine them the same way you did with the first batch.
Offenders you might start to notice now: -pacing. It’s one of the more complex problems because it’s really hard to examine and there’s no real rules. But you’ll see it with that outsiders perspective; sometimes you just spend way too long on some things and way too little on others.
-description. This is super personal, some authors rarely describe more than the literal events of the story, and occasionally mannerism. Some authors go super super hard on describing environments, what characters look like, how things make characters feel. What’s more important here is feeling out a ceiling and floor. Fall below the floor and you can’t imagine the scene or setting at all. Rise over the ceiling and the pacing and flow will tank, the audience will be bored to death.
-consistency and flow. Inconsistencies between sections of the story and sudden jumps that don’t feel precipitated will immediately yank the audience out of the story, and they can be tricky to avoid because ~you~ know where the story is going but the audience won’t.
At this stage, the trick you’ll probably want to learn is to summon the outsider perspective on demand. regularly. At different stages of writing. Honestly this worked out best for me by literally inventing a character in my head named Joe Averageguy. He has a dopey voice. And I “ask” him what he thinks frequently while I’m writing. This has many downsides that you can probably guess but it does solve the problem I had with summoning that audience perspective.
After this? Well, you’ve probably been writing for like a year, hopefully with some consistency. Push yourself out of the nest however you see fit. Pursue a significantly longer story, have someone you know read some of your stuff, shut you could realistically publish something with how open that process is now.
Also: consider having an LLM critically examine a story or two (if you can get it in the context window) KNOWING FIRST that your story now belongs to OpenAI or whoever. This approach still has real value; it’s one of the only things those LLMs are consistently good at and it is nearly immediate reasonable feedback. And that is going to be HARD to find. Don’t just say “critically examine this”, process, and bail. Probe it with many questions like similar authors or target audience information. When possible, modify the LLM to not be a sycophantic worm. Just never let it feed you direct phrases or sentences. All LLMs have a firstly distinct voice, and that voice sucks. Don’t let it inject your writing or your brain with its bland corporate filth.