This title isn't from me; it's from an early access postmortem of the game Book of Demons that you can read here.
I don't like this question. Different indie devs getting similar results might come to different conclusions and it's really hard to get people to read between the lines to see the real picture. Book of Demons sold about 7,500 copies in 2 months in early access, which is similar to what my last game March of the Living achieved as a full release in 2 months (around 8,600). I was happy with those sales, but apparently the developers of Book of Demons came to a different conclusion.
The difference? I spent 6 months on MotL while Book of Demons has been in production for 3 years. I'm the only dev behind MotL vs 7 core team members for Book of Demons. As you can see, I didn't need to sell 35,000 copies to call this project a success and I never expected such result either (all indie devs secretly wish for such success, of course, but eh....)
So asking the question "Why is selling good games so hard?" here seems to say that either MotL didn't sell well or that it's not part of the "good games" category so it sold as much as it could considering this. The truth however is probably that Book of Demons and MotL are selling as well as should be expected. The problem is that you can't use development cost as a premise for the question "Why is selling good games so hard?"
Two things from the postmortem caught my attention in a weird way:
"To recoup the costs, we need to sell about 35 thousand copies of the game."
There's nothing wrong about working on a big project and trying to get big results. But you need to accept that things may not go the way you want when you need to sell 35,000 copies just to break even. It might seem like a low number when compared to top sellers, but the majority of games on Steam will never, ever sell that many copies. And this brings me to the other point:
"In theory, selling this much should be pretty easy. After all, hack & slash is a very popular genre, and games such as Grim Dawn, Torchlight, Path of Exile, Van Helsing are able to move hundreds of thousands or even millions of copies."
You really can't believe such thing. This is not a business plan. Maybe it wasn't meant as an actual statement of expectation but it's not the first time I hear such thing. If it were really as simple as checking what sells a lot and copy/paste results we'd all be doing this. I can't look at the success of Firewatch and then start working on my own walking sim while telling myself "well, that game was a hit so there's a reasonable chance mine will be as well". At best it might hint that people are ready to buy the right type of game of that genre (something unknown for the first walking sim released) but you can't put a number to that conclusion.
Game development isn't an investment, it's risk management. There's a good reason big studios are pumping sequels after sequels; it's way less risky than working on a new IP. This is a vital lesson that indie developers must heed, even if we want to work on exciting new projects instead of yet another sequel.
It might be a boring thing to say, but working on the game you want to make and hoping results will be proportional to your passion is simply blinding yourself to reality. It's like the advice, "don't quit your day job." You should be prepared to the fact that things might not go the way you dreamed they'd go. You can get another mortgage on your house but, if you do, you shouldn't be surprised if you eventually lose it. Going big is never a guarantee of anything.
My friend Brian ‘Psychochild’ Green wrote about reasonable schedules with another game, Brigador.
I wish the best to Thing Trunk and since their game is in early access it's reasonable to think that the current results are not the end of the story (for example, as a dev myself I don't buy games in early access as I have my own projects to debug and test so I really don't feel doing so for other games).
We need to be careful though when reading postmortems and trying to interpret results. I know quite a few solo devs that would be thrilled to sell so "few" copies in 2 months but the press still probably wouldn't run stories about how they succeeded. The risk devs take can't be considered to define what are normal and expected results. If my last game would have sold 35,000 copies, I could literally do nothing for the next 3 years and there's nothing quite normal about this.
The successful indie devs I know are very careful about scaling up as they know that the top sellers list on Steam is an anomaly and not the norm. There's nothing wrong with taking bigger risks but it doesn't mean the reward at the end will also be bigger.