Star Citizen's business model is so different to just about any other mainstream PC game that it can't really be used as an example of anything.For a start, in terms of copies sold they've sold under half what Elite Dangerous has. ED is not trying to be a game funded primarily by a tiny number of people putting in 4- or 5-figure sums; even in the Kickstarter era it was better balanced than that!Sure, but that's not really how it works, is it? After all, we have the perfect counter-example in Star Citizen. They've clearly shown that the more you promise the more you can earn - as long as you don't clearly demonstrate incompetence.
And how much "respect and trust" do you think Frontier would be generating right now if we were being assured that "the next era" was still in development and was definitely coming soon, and we were all still running 3.8 with no CGs, no Galnet, no Thargoid war, no Powerplay rewrite, no new ships, etc. etc. while they were "looking to the long term" and "making sure they got it right"? (So, basically, everyone gets treated how the consoles currently are, but starting a couple of years earlier)In that light, focusing on the short-term cost of a project going over-budget is literally the last thing in the world you should be doing. A one-time-cost can be earned back eventually, but respect and trust, once lost, is virtually impossible to fully restore, and priceless to maintain.
There are costs in both finance and reputation to not releasing content too.
I'm not saying that releasing Odyssey when they did was a good decision. But the point where they could still have made a good decision was probably back in 2018 when they were scoping it in the first place, by setting out to do something they could realistically achieve. By the time it got to late 2020 they were just picking between various different expensively bad options. (Of course, if Frontier was in the habit of setting out to do the realistic, they'd have taken one look at the concept of Elite Dangerous and decided to start with Planet Coaster instead, so there's a balance to be drawn there from our point of view)
The other thing is: if your project is a year late, with the corresponding budget overrun, and still not finished, how sure are you that your estimate that it'll definitely be done if you have another six months is right? You shouldn't be, at that stage, and neither should the people offering the budget. Yes, a one-time cost can be earned back eventually - but only if it's actually a one-time cost. If your general practice is for "2 year plans" to actually take five years to complete - and Odyssey was neither the first nor the last Elite Dangerous content to have a massive schedule overrun - and then to take fifteen years to pay off that cost and become profitable, then by the time the first one is profitable you've got at least three more making a loss to cover, if your product and company lasts that long at all. "Eventually profitable" doesn't pay the bills, though whoever buys up the back catalogue rights from the wreckage will have a better time of it.