A premise of the Internet publishing revolutionaries has been that because distribution is ubiquitous, the marginal cost of content approaches zero, meaning that information should become much cheaper than it once was. In fact, it should be free once a modicum of cost is paid for — hence, the open access model, for instance, or the devalued online access prices for institutional access, for another instance.

You still hear this argument, even though real-world economics seem to defy the logic. Just this year, Clay Shirky invoked the “marginal cost” argument in his most recent book, “Cognitive Surplus,” stating that, “Information can now be made globally available, in an unlimited number of perfect copies, at zero marginal cost.”

However, for those of us inside the belly of the beast, just the opposite seems to be happening. That is, digital publishing seems to be getting more expensive while we wring costs out of print as we draw it down.

One problem seems to be in the model we’re using to describe the costs involved with digital publishing, relying on a manufacturing model rather than a more suitable model that involves sustaining and spreading fixed costs across a longer time horizon — that is, the software model or, as I’ll explore later, the model being used by digital distributors.

The software model arrives at pricing by spreading the costs (and profit expectations) of a product across its anticipated customer base for a particular release and support timeframe. If we begin to think of content as software — issues as releases, support as provision and maintenance — and have a relatively stable customer base, we can begin pricing online properly. We just need to allocate the costs properly.

Because we’re using the wrong model, we can confuse fixed costs with sunk costs, marginal costs with average costs — all while not taking into account the time horizon we’re after for maintaining access to digital information, the trajectory of improvement and diversification required in a flourishing digital ecosystem, and the cost of human talent.

One part of the digital economy has gotten the model right — distribution providers.

Distribution has become a major expense for consumers — I talked about this earlier this year, showing how spending on Internet access alone is now double what people were paying for information in 1999. And this doesn’t even include things like cell phone bills, text messaging charges, or the price of devices.

It’s been noted that the high prices of distribution hammer societies unfairly. For instance, a recent article in the Guardian argues that libraries remain important because they provide Internet access to economically disadvantaged members of society. By helping to bridge the digital divide, libraries give the poor access to information they otherwise couldn’t get.

The digital distribution business model takes fixed costs — the installation of a satellite system, cellular towers, fiber optic cables, and network switches — and charges both what the market will bear (a lot, it seems) and what the provider needs in order to recover the investments made in the infrastructure, while also making enough for ongoing improvements, enhancements, and maintenance.

This is much more like the economics of digital publishing than the manufacturing model is — first-copy costs plus run-on costs just don’t make sense in the digital world.

Instead, we’re entering a realm of high fixed costs that have to be spread over a longer timeframe, which must be recouped while new charges are accumulating from maintenance and enhancement activities. And as editorial and management costs migrate from a hybrid print/online budget (in many organizations, costs are still presumptively print), the fixed costs for digital will only increase.

Then, publishers, authors, and retailers will have to absorb the reality of a digital-dominant model. My guess is that pricing will increase dramatically across the board as the fixed costs, talent costs, and full-on business expenses have to be met by digital business.

Interestingly, music pricing — one of the more mature areas of digital business — has been changing as distribution has started to dominate the digital side of the business. Despite a heated battle between Amazon and Apple over pricing, Apple’s per-song rate continues to creep upward, with many songs — especially more popular ones — now selling for $1.29 each, essentially a 30% price increase. This shows the power a distributor (which is all Apple is in this game) can wield when they create a full distribution value-chain (files > store > e-commerce > device) that dominates. Knowing their costs and having the right model (the distributor model), these business know how and when to raise prices.

Condé Nast is looking to end the era of $12 magazine subscriptions, basically reflecting on how well distributors have managed to eke value out of the market:

[Their CEO] expressed confidence that the consumer would come around, noting, “They pay $180 a month for a cable bill.”

Publishers need to move beyond the print model of sunk costs and marginal costs. Digital publishing presents a different economic game, and defraying fixed costs while paying for talent, infrastructure improvements, and ongoing publishing takes a model more like software or online distribution.

Otherwise, we’ll all be marginal and sunk.