D:OS2 was also present on Kickstarter, which is much more prominent than Fig, giving it a leg up in terms of "marginal" backers.
So was the original game. What's significant about their campaign is that the second game had double the amount of pledges as the first. Obsidian chose a different platform so their results can't be directly compared, but I would still question whether Pillars of Eternity 2 could've gotten twice the amount of pledges as the first.
They have more than one hundred people on payroll. Do you think that will gamble with their confidence by making their dream game? Especially now that they sold a lot by delivering a mediocre game? No. They will try to do the same thing, but more efficiently. They think like a publisher. This is their job. What them want have nothing to do with it. It’s checklist box design all the way. We are talking about people who made a risky choice when they're younger and didn’t know better and now are trapped in a poisonous environment. Tim Cain is just a shadow of his former self. His passion for cRPGs is long gone.
There's always a risk, no matter what publishers tell you about market check lists and safe investments. Even Electronic Arts ends up with many commercial failures, and their way of doing business is losing steam:
https://seekingalpha.com/article/4099180-electronic-arts-losing-grip-video-gaming-industry
A large fraction of the most successful franchises in recent years come from commercialization of custom maps and modifications, the design of which came from
players, not marketing committees. Warcraft custom maps gave birth to Dota and League of Legends. Team Fortress gave rise to Overwatch. Horizon: Zero Dawn was described as Guerrilla Games' most risky pitch. Even Counter Strike started off as a modification. Sure, there's still safe plays like Call of Duty, FIFA, and World of Warcraft. But the industry is no longer ruled by them.
But any market is small in comparison with Triple-A. The proper comparison is between BG, and the rest. If you consider that, the BG series was never small. In fact, alongside with Diablo, it was always on its own level of popularity turbocharged by D&D franchise. PoE is practically a skeleton of a game and they sold almost 1 mill units with a ridiculous price tag. That’s a lot. And it’s not just Obsidian, mind you. Harebrained Schemes delivered sub-par games as well, and sold a lot of units too. cRPG players like to think of themselves as hot shit, but they are not. They are mediocre and predictable. The mere resemblance of a traditional cRPG with checklist design is enough to satisfy their tastes. Anything beyond that it is risky and will only get you in trouble. If they weren’t, we would have more developers making risks.
While I agree with you that CRPG fans are mediocre and predictable, that's not the issue I'm talking about. You can make check list games, as long as the people you put in charge of them, actually like them for what they are. The problem I have with putting Josh Sawyer in charge of these "traditional" CRPG clones is that I don't feel he actually likes them. Reading his posts about CRPG design, it's quite obvious that he feels traditional CRPGs are broken; yet, as we've seen with Pillars of Eternity's release, his changes are not what the people who enjoy these games necessarily want.
That's why I'd suggest Obsidian put him on a game he'd enjoy developing, and be passionate about. You might argue it's a risk, but so is having him develop games he doesn't enjoy. Leave the check list games to the people who love them.