nobody seems to get it...
you don't want to sell games to people, you want people to subscribe to your games!
It's a big strech to suggest 'nobody' seems to get it. I'm from the MMO world, where people understand subscriptions. But a bird in the hand is worth two in the bush, and you need to sell your current game now, not your next one. When it takes you years to make them, that's the main concern.
not having to process returns? less bad press? no thinking "oh! we're rich!", only to discover that have to give a lot of it back?
Distributors process their returns. There wasn't much bad press, just whining in reviews. And the amount of money refunded (not given back - they won't have even seen it in their account, given how distributors work) is very small compared to the overall revenue.
demos can make use of a limited subset of assets. with the rest of the assets not included
You missed my point. It's a procedurally generated game. Most of the assets are in the code.
sounds like they just didn't make a good demo.
The original Deus Ex demo was bad. But that's an extreme outlier where the demo was bad but the game was great. Normally the gap in quality is much closer; and all the demo serves to do is turn off some potential buyers who were originally on the fence. The big companies have done the numbers on this and they're very sure about how it works.
It may well be different for a lesser-known game, where you need the publicity. But well-publicised games aren't in that position.
a solid reputation is critical to long term success in any industry.
In an ideal world, every game we ship is exactly what we wanted it to be; but that's not always possible. So some games will be a bit disappointing. Releasing demos won't make the end game any better, but it does make the developer more likely to go out of business. There might be some less disappointed players as a result, but I don't think that's a price worth paying.