Hm. It takes one person about 4 weeks to do one book? Would you say that's fair? Besides the LW staff, you're probably the one with the most experience in building data sets...
So you've got paid staff for that month, plus infrastructure/overhead costs. How much is that? How many people purchase said content? Is it enough to cover costs? And remember, you want to recoup your costs at least within the same fiscal year, probably within a few months of release. Anything after that is "gravy", so to speak.
LW is not a public company so there's no way to know how those numbers work out. Therefore, it's not possible to even guess at whether the margin is large or small, and it probably varies for each Paizo publication.
I would not want to be the middle manager responsible for making sure that such development is and stays profitable.
That depends. A common technique nowadays is to outsource your server processing to a cloud service like Amazon Web Services. There are tiered plans where the customer plays a flat amount, then incrementally based on usage. If usage is low, there's a minimum charge and when usage spikes up, costs go up. There are some big advantages — automatically capacity increases, for example.
But I haven't been paying any attention to any of that, so I'm wholly uninformed regarding how LW is planning to do manage their server side requirements...
