If a game thrives or fails, starts when choosing the engine. From a technical perspective, RPGM isn't the best choice. As an anecdote, devs like Redikal and TwistedScarlett regretted picking that engine due to technical difficulties later. In terms of inclusivity, RPGM was created with Windows in mind and not for Mac- or Linux-based systems, unlike Renpy. Yes, some people prefer to play their h-games on these OSes. When selecting RPGM you artificially limit yourself, which—from an economic perspective—is stupid if you're new to the scene and desperately need funding. Moreover, letting the
unique fetishes run wild was Dokan's second-biggest problem. They should've picked and limited themselves to a handful of fetishes, where there's a demand for and that fit the game.
Quality over Quantity! Most people like their Vanilla stuff and aren't necessarily into NTR, Femdom, or hairjobs. Try to cater to too many unique fetishes, and you'll be in development hell. Regarding the
voice-acting. Never implement VA from the start because stuff will eventually change during development. Finish the game first; afterward, you can still decide whether you want to implement it or not. Furthermore, voice-acting is nice to have but isn't a necessity. Most voice actresses are unreliable. There's a considerable demand for ASMRtists, and they'll primarily work for whoever pays more. It doesn't shock me that Dokan's VAs started ghosting him shortly after. They got paid, someone probably made a better offer, and they decided it wasn't worth their time working with Dokan. My two cents. Sorry for the essay,
LoveHurts.
EDIT: Remember that voice-acting is expensive, especially if you're at the start of your dev career, don't have a substantial financial budget and just started accumulating fans who monetarily support you.