I second the importance of adding ways to assess observation quality. For me, my main motivation for doing IDs is to learn more about the taxa I am identifying instead of curating species range maps or helping other people learn how to do IDs. Only high quality observations showing key characteristics are useful for that purpose. I tried IDing regardless of quality before but got burnt out and frustrated. Now I switched my interest to microbes and the fact that you need a microscope to even make an observation of most of them acts as a good quality filter in and of itself and it’s a much more enjoyable experience. The main problem is now that I’m still pretty clueless about how to ID microbes instead of having inadequate details to work with.
The best approach I found to focusing on the higher quality observations was to only ID obs from a list of users that I found who tend to make good observations. And I am very conscientious about the quality of my own observations as well. Right now there is basically just one person IDing my stuff, often coming down to subspecies when I couldn’t even guess at kingdom. So I am very paranoid about not filling my feed up with just any old observations in case it causes me to get unsubscribed. I find that posting observations as a small constant trickle instead of in huge batches seems to get me higher ID rates. I can certainly relate as when I am IDing stuff and see a huge batch of things from one user I tend to start skipping them. So maybe adding an option to spread out when your observations get published would be a good addition to the draft mode if it ever gets made, as well as an option to just present a set amount of observations for you to ID each day.
Another problem is when someone posts an observation with excellent photos showing all the key characteristics and ID’d to species and they still tend to sit there without any additional IDs. I tend to glance past these observations since I usually don’t know how to ID things to species, but if I would just pay more attention to them I would eventually learn how to ID them. Right now, there is a notion that you should generally avoid adding coarser non-disagreeing IDs to observations, such as #3 in Identification Etiquette on iNaturalist - Wiki. I think that guideline should be chucked out the window because adding a coarser ID would at least cause me, and presumably others, to pay more attention to these sorts of observations.