3D printing is still fiddly, complex, error-prone, expensive and slow.
This. We could stop here but let's not, I have more to say as well.
FDM style printers (the cheapest kind) require wrapping your head around calibration, nozzle diameters, temperatures, slices, alignments, supports, bed heating, the properties of PLA / ABS and all the rest. If you're lucky you'll set the printer going and hours later your efforts will yield some crudely finished single colour part. If you're unlucky you'll come back to discover something that has skewed left, warped on its base, or turned into some dante-esque spider's web that has stuck to everything.
#1 thing that should have been done/should be done to improve 3d printer uptake is working automatic bed leveling. AFAICT (and I am about to test this theory) the best way to accomplish this is with an inductive sensor designed for iron, detecting an aluminum print bed. Aluminum is pretty much the best base anyway, since glass can break all to crap while you are removing your print job. Inductive sensors literally cost five bucks and the hardware needed to interface one which isn't already present on e.g. RAMPS literally amounts to a grand total of one voltage divider, aka two resistors. IIRC 10k and 15k are typical, but whatever it takes to get the output signal down to 5v. It doesn't even matter whether you get one which is NO or NC because the software (Marlin, in my case) can detect either type of signal with correct configuration.
The #2 thing that should have been done is more centralized sharing of print temperature settings for materials. After bed leveling, just finding working print settings is the biggest PITA. If you're not buying from someone who gives you some starting values to work with, then you have to go on a web safari quest to figure out where to start.
Way down the list, probably, is eliminating all these stupid microswitches. Hall sensors are more reliable and cost little more. Even if you have to get them on eBay, whole PCBs with a sensor mounted on them are only about twice as expensive as microswitches worth buying.
As for the single-material print job issue, this is the printer I've got, it's reasonably priced and its great sin seems to be a crap auto-leveler. I've got an inductive one in the mail. Deltas take most of the fiddly-ness out of the system at the cost of potentially dropping a hot extruder on your print job if you don't program them correctly. Which all comes back to your point... they should be set up better for the user. I am willing to do all this fooling around to get the thing working but the average user just wants to print and have something come out.
I actually don't think having a crappy-looking single-color print job come out is what keeps most people uninterested in 3d printing at all, though. I think it's really all about the PITA factor. If you offered people a machine which was both cheap and easy to use, they'd jump all over it even if it only had typical resolution and speed and was limited to a couple of materials.