I was recently lucky / cursed enough to see a live performance by infamous electronic music act Autechre - they of "David Lynch likes them" and "at some point there was allegedly a possibility of them doing the soundtrack for Metroid Prime"- fame.
They're an experimental duo who have been around since the early 90's - they sound like this…
Their music stems heavily from process - they're big fans of the modular audio programming environment MAX / MSP (in contrast to musical tools used by normal people) and tend to focus on generative techniques, self-consciously FM-synth farting on various conventions with subversive glee. Like the Nathan Barley tiny hats and glassless glasses worn by their fanbase, it's all a bit silly in that microdosey, Pitchforky, subreddity, art-installationy way. But there's also something wonderful about it - sitting in the dark listening to immense architectural sweeps of sound that nobody in their right mind would put next to each other is an arresting and evocative experience.
I'm generally an advocate for understanding your audience and using that understanding as a touchstone, corollary and anchor point for creative work. I'm sure that Autechre - in their clattering, screeching, throbbing madness - do some of that. But sometimes it's also good to just let loose and do whatever you want.
If you are going to just go for it and follow an impulse wherever it takes you, then you should really commit. Proactive creative aggression, cohesive aesthetics, backing up your surface level ideas with real substance and finding a viable distribution method to connect all that with actual human beings are critical. I've talked this through with a number of game devs who've found themselves lost after having a big commercial success - if you have the freedom to explore then why not take the opportunity? Focus internally - on processes you enjoy and outcomes you appreciate - then see what happens.
It's important to understand that this is a high-risk / high-reward strategy in terms of commercial viability. You probably wouldn't ever get to Kenshi or Undertale via a more conventional product design approach, but you obviously run the risk of just alienating everyone who isn't you. You might be Autechre, get signed by Warp Records and spend your career doing exactly what you want with the blessing of a powerful marketing and branding machine…or you might be stuck in your basement forever. There's a lot of luck involved.
Another danger is ending up in the middle - pretending to yourself and others that you're breaking new ground creatively when really you're just drifting in No Man's Land. That's a bleak zone where your work won't be attention-grabbing enough to matter to anyone, and it won't appeal to more conservative consumers who need some conventional handholds to grab.
My plea to you if you're considering flying in the face of convention is to do it consciously. You need to be comfortable with complete commercial failure and create an acceptable context for that - dragging others with you into the void's cold embrace might be something you regret.
If that all sounds a bit terrifying, then try to fit in a "B Project" - something small and low-risk in terms of time and money which allows you maximal freedom: that'll give you a chance to test the water without cilff-diving into peril.
I've talked about more audience-centric approaches before, and I'll probably revisit that topic soon. That path lends some constraint and definition to your experiments - it can be very fruitful and perhaps lead to more consistent success - but don't be afraid to buck the trends from time to time - you might end up somewhere wonderful.
Thanks for reading - I’m hopefully going to be doing more of this soon so do subscribe if that’s your sort of thing.