You will have noticed, if you're thinking about buying some server-side software, that it can be really hard to get a straight answer from the salesperson on how much it's going to cost. Antarctica is no different, but I'm pretty well convinced that this is the right way to behave.
Vendors have historically used a bunch of different pricing formulas: per-seat, per-simultaneous-user, per-number-of-objects, per-transaction, and so on. The trouble with this is that all of them without exception break down in the real world.
To see why this is, let's think of software purchase in simple old-fashioned terms. Someone who's buying something wants to get enough benefit from it to make the investment good. By analogy: anybody would spend $10 to fix a $1,000 business problem, but nobody would spend $2,000, and you'd think pretty seriously before you signed up to spend $750.
The problem is that the business benefit you get from the software might not have anything to do with the number of objects or the number of transactions or any of those other measures, and the conventional formulas might produce really silly prices.
So a good software vendor, I think, is going to start out by asking a whole bunch of questions, trying to figure out what the customer's problem is, and how big it is. If it's a big problem, and one the software is really going to help with, the chances of a happy ending are good. If it isn't that big, possibly we the vendors should politely suggest that maybe it really isn't worth throwing server-software-scale money at.
For example, our first ever deployment, which worked really well and made the customer very happy, covered a grand total of less than ten thousand records, with never more than one or two people looking at it in parallel.
Another situation we're working on now is a relatively small company with an immense database: tens of millions of records, and hundreds of thousands of users.
In the first case, conventional price-book pricing would have produced something so low that it wouldn't have been worth our while to try to make the sale. In the second, the list price would have been so high (north of a million dollars) as to send the customer running screaming in the other direction. In both cases, when they asked for a price, we (politely) dodged the question and came back with a bunch of our own questions about what the problem was. And in both cases we got a really good outcome.
