Yes, but I’m not an artist 
It’s also not a matter of making one character with a different look; it’s a matter of building mix-and-match “paper doll” or “avatar” characters, that can change their clothes and equipment at runtime. Having built this kind of system from scratch and worked on similar systems at other companies, including for UGC driven companies, I’ve found that many artists would rather build something custom and eye candy, than sweat to match a bunch of “arbitrary” technical specifications, but if you can actually make it all match up, it can be great, and very powerful for those who go that route! And if you have the right templates and easy-to-use verification tools available for artists, you can get communities of artists who build particular things they’re good at (shoes, or party dresses, or hairstyles, or whatever) and sell them in the marketplace, and then mix-and-match for your character builder.
But it doesn’t! Re-targeting may make it possible for me to buy the Elf Swordswoman With Improbably Skimpy Armor and re-use some animations I already have, but it does pretty poorly with “make fingers match up to grips” and it doesn’t let me take the hauberk from that Well Protected Orc character and cover the most vulnerable parts of the Elf without significant custom surgery – and it doesn’t help any at all at runtime. Consider any avatar-based game you can think of – MMOs, open world, sandbox, 3D chats – that’s the kind of “runtime mix and match” that’s missing.
There are a few “character customization systems” on the marketplace (I think I may own all of them?) but they don’t come with any real variety of clothing (or any at all, in some cases – what does “customization” even mean there?) and there’s no extra parts to buy for them. The “advanced” ones seem to go overboard on morph targets – sure, that’s fun to build and also fun to use, but misses the real need for “mix and match humans.” Honestly, some of the better ones are the Daemon3d girl/boy sets, but even those are only a part of the way there. (And then they’re a bit cartoonish, rather than “default realism-ish” – there’s a certain question about art styles not matching between parts, too … but they’re still among the most varied!)
Death by aiming too high, perhaps?
I would be super happy if I could get five each of “low fantasy” male hairstyles, underwear, upper body clothing, lower body clothing, and shoes, to all match up, and then the six or seven major skin tone textures to go with it. Even better if there were well defined body shape morphs going off “normal” towards “heavy,” “muscled,” and “thin.” Even better better if there were texture set morphs to match those additional shapes, too (muscle tone, especially!) If I could get the same for females, that’d be even better. If there were about six major morph targets for the face with well-defined semantics and names, that’d also be swell (maybe some well-defined morphs to go with the well-defined skin tones …) Add expansion packs for modern-street, and sci-fi, with similar assets for the upper/lower/shoes parts, and it would probably do very well on the marketplace. I mean, I doubt the marketplace is a major income source for most sellers anyway, but the “my open world needs customizable characters” and “my RPG needs different kinds of armor” and “my virtual changing room needs wardrobes” needs would be met.
I have worked on, led, and in one case built from scratch, three major shipping avatar systems that support this level of modularity, starting in 2001 and ending in 2017. It can absolutely be done. It can even be done with user-generated-content as the driver. It needs strong standards, it needs good support assets and tools, and it needs a big enough foundation of assets to give it critical mass. And some artists would not want to build to this spec – that’s totally OK, as long as some would!
Building just the foundational assets from scratch would cost several hundred thousand dollars, at least, assuming off-shore outsourcing rates. Again, depending on the level of sophistication required. (I’m thinking 500k vertices high-res, baked to 35k vertices in-game, perhaps?) Meshes built for this system would be slightly higher-poly than “optimized” because you’d have cuts along well-defined paths for “when wearing shorts, these skin bits go away” and “when wearing short boots, these skin bits go away” and so on. If I were independently wealthy, I’d totally throw money and time at this! Someone would build the runtime system for all of this once, to make it easy to get going (I’d be happy to) and push it all out, and hopefully there will be enough critical mass to actually establish something for the rest of the community to build off of.
It’s really that, though – critical mass. I don’t think creators, on their own, would achieve that. I would love to be proven wrong, though 