I recently picked up the Early Access game "Prison Architect" on Steam. And I was very impressed. I tend to really enjoy these types of games, mostly due to the idea of thinking tactically and allowing the AI of your "minions" to enact your grand plan. However I was struck by an interesting realization.
So a quick disclaimer: I know that this game is still technically in "alpha" at least as far as the developers are still working on it.
The beginning of the game now includes a short tutorial (with great art btw) that involves teaching the player how to hook up power to an electric chair. It does a great job at teaching the user how to handle utilities that aren't visible in the normal mode. However the game's current objectives don't include goals that involve an electric chair. Instead it seems their latest update involved adding logic gates to control doors, cameras and other automated things. It adds an incredible amount of complexity that most players won't even need to know about. I think this is because the active community demanded such things.
This to me shows that for small game teams, appeasing your active audience could be more important than making a game easier for new players. With the small games portion of the industry exploding, "custom ordered" or niche games may become the best way for small teams to get a dedicated audience for their products. It can change a fan into an advertiser of your game so that when it does go "live" they will be much more interested in spreading the word to show the developers support.