While conceptually I agree with the premise here, given that most peoples' desires in creating Fire Emblem-based projects stems from having played through one or more of the official games, I don't know if it's quite easy for people wanting to delve into this to not immediately jump to crafting a long winded epic like what they've already experienced. I mean, for some people, that's part of the reason that they play the games in the first place - not necessarily the overall story (because, let's face it, they're only sitting at about "just okay" on the scale), but the combination of the characters, their development, and how those intertwine with the plot and other characters. (Also, don, you forgot about all of the people that want to dive right in and make cliche hack #467 (Scouring, Black Fang prequel, Black Fang sequel, crossovers / reskins, etc.) - I'd say that they're overall more prevalent that people that want to build their own world from the get-go.)
In regards to the bullet points in the opening post (before I delve into my personal suggestion), I think the last two should be of special merit. Some other options for those two points which I didn't see further discussed would be to craft an experience where you're already in control of stronger units (high tens and low promoted) as opposed to even something as low as Level 5. When you have stronger units with built up stats and weapon levels, it can open up more possibilities for unit loadouts and, consequently, greater enemy diversity in a quicker buildup across the much shorter campaign. Especially if the author is dictating what the possible loadouts are (i.e. no shops, limited weapon drops or chests) and what enemies the group would be thrown up against.
Furthermore, I think you can kind of replicate the feeling of a complex class tree by making the characters and classes much more unique and less "generic" than their standard counterparts - as you're building something specifically small, you can repurpose swaths of lower-tier classes that aren't being used into giving each individual character a personal class of sorts or trying new things instead of just going with the stock class types. And, I'm not even referring to something that would involve large amounts of work (i.e. making new animations for new classes or upon adding weapons to old classes) - consider just how much could be changed by tweaking the roles of the standard classes and converting them into something different. Take for consideration converting the Swordmaster class into (conceptually) a class that's adept at skillfully blocking/parrying attacks with its sword - change the class caps of Swordmaster (M) to 60 HP, 23 Str, 30 Skl, 24 Spd, 28 Def, & 22 Res (though I suppose that's conceptually close to a Hero, albeit mono-weapon...). Or, convert the Druid into the Mage's promotion and give it only Anima and Dark (and, if you're hacking FE8, throw it on top of one of the three characters that can summon and give it that too). There's several possibilities to spice things up, especially when you have a much smaller cast to focus on and a smaller scope to plan for instead of making sure all of your ducks are in a row over the long haul of a full-blown campaign.
That being said, my overall suggestion would be that if you're world building your own story and setting, craft a small corner of it to set up as a demo / proof-of-concept / introduction to your world that sows the seeds of the larger project you inevitably want to tackle. Heck, plan out a couple of them - they don't need to be actual short stories or analogues of them, just something "canon" to your setting to get your feet wet. I mean, Yeti said it already, that's kinda what works with his Trial Maps (although, the specific purpose between this and those would be different since his were ostensibly to code mechanics into the engine and then test them, not expressly to world build, but that happened as a result); all you'd be doing would be forming a small-ish, cohesive tale with each chapter carrying over to the next like a regular FE (unless you've specifically crafted something where you don't want that to be the case) with a known, early end-point that in some way sets up what you want to make in full.
Heck, half of the ROM hacks that I've actually found fun and enjoyable were things that were like 1-2 chapters long and were solely challenge map / chapter mini-looks at things the creators were working on, simply because they focus on one small set of things, go relatively wild with character distribution, new items, etc. and let you play around and experiment/explore what they've made.