> The main schedule constraint is that marketing spend must be paid months in advance for the purpose of reserving specific time windows. That spend is often equal or greater than the development budget. And, if you miss your time window, too fucking bad. You will have no marketing, no more money to buy in again and your game is now a flaming money hole.
That's a good argument for being date-driven. However it strikes me that the best way to deal with that is to have a disciplined project management process so that you have a good estimate for when product development will actually be done.
That's the theory. And, everybody claims that's what they do. But, in practice it is exceedingly difficult to estimate not just software, but software that is premised on delivering novelty, creativity and a volume of original and highly-technical art greatly exceeding the budget of the engineering team.
On top of that, there is pretty much no defined path for learning how to be a game production manager. People come in, in myriad ways, pretty much completely unskilled as junior producers and mostly learn by following their seniors. As a result, project management skills in the industry are generally quite... poor.
> That's the theory. And, everybody claims that's what they do. But, in practice it is exceedingly difficult to estimate not just software, but software that is premised on delivering novelty, creativity and a volume of original and highly-technical art greatly exceeding the budget of the engineering team.
I hear ya. I don't have a background in game-dev, but I have done some tech R&D, and worked with others who have a lot more such experience, and so I understand the challenges of setting a delivery date when the problem is so open-ended. It's helped when you can spend some time upfront trying to eliminate the biggest sources of risk via prototyping, etc. Would that be called something like "pre-production" in game-dev?
However, and I don't mean to be glib here, but it seems to me that most creative software development isn't so revolutionary that it can't be estimated (or, at least isn't so radical that the feasibility of given due date cannot be determined) by a team with deep domain experience. I suspect, though cannot prove, that the fact the industry tends to burn people out after only 3 years is largely what causes the burnout. If they did a better job at retaining experienced engineers the need for crunch mode might diminish greatly.
> On top of that, there is pretty much no defined path for learning how to be a game production manager. People come in, in myriad ways, pretty much completely unskilled as junior producers and mostly learn by following their seniors. As a result, project management skills in the industry are generally quite... poor.
I think that's probably true for the software industry more generally. It's no knock against the people who are project managers, I just don't think there's often clear guidance for what the project manager is really supposed to do.
That's a good argument for being date-driven. However it strikes me that the best way to deal with that is to have a disciplined project management process so that you have a good estimate for when product development will actually be done.