You're right,
And essentially the only way to achieve the optimum (maximum price discrimination?) is to know that specific consumer's demand curve -- which is also the whole idea behind "contact us" pricing pages.
I also think there is a limit to how complicated you can make the payment system before the user just gets confused. So that definitely should be considered. But two-part seems to be a familiar model for most people.
There are ways to find analogs for their demand, particularly if they are using your software directly to make money.
For my company, we try to target our monthly pricing to be between 1-2% of what our estimated costumers revenue is for the same period.
This method only works if use of your software is directly tied to the production of their product. It falls apart if your software is merely ancillary.
the absolute optimum to capture maximum consumer surplus would be infinite-part pricing, i.e., charging each customer their maximum willingness-to-pay price. but it's very difficult, if not impossible, to determine that price in an efficient way.
why not three-part (or higher) pricing? because you increase the complexity of the buying decision without enough of a payout.