I’m certainly sympathetic to this idea. The marginal cost to serve an HTTP request is vanishingly small, but the fixed costs to develop the application itself (

I’m certainly sympathetic to this idea. The marginal cost to serve an HTTP request is vanishingly small, but the fixed costs to develop the application itself (mostly labor) is pretty high. That means your cost per user is pretty high at first, since all that work supports a handful of people. You seek more users to amortize costs further, and it works until you start accumulating enough HTTP requests and user data that costs start to climb again. And of course more users means more use cases and user agents, which require ongoing maintenance investments. The whole process doesn’t have to be expensive, but it’s certainly not free. You can build very cool stuff and give it away for free, but sustainability and scalability ultimately require revenue. The magic thing about a successful business is the ability to cover execution costs, support the development team, and still leave value on the table for users. I think digital services have drastically different economics depending on (1) how adding a NEW user changes the value proposition for EXISTING users, and (2) something like the user’s start up/discovery cost relative to lifetime value A direct payment model makes sense when your users’ value is independent from the size of the user base (assuming performance scales at least linearly). These services can tolerate moderate startup/discovery costs. For critical enterprise services, startup costs can be high because the lifetime value is still much larger. If value scales with the user base, as in a social network like good reads, then startup/discovery cost must be pushed to zero to grow the user base as quickly as possible. A paywall slows user base growth and reduces value for those users that actually choose to pay. So far, advertising is the only known way to monetize (and thus sustain) a digital service while maintaining near-zero startup/discovery costs for individual users. Micropayments, even with good UX and low fees, increase joining costs relative to advertising. Thus they will reduce the value of the service to paying users if value scales with user base size, but would benefit services where value is independent of user base size. Federation is maybe the best way out of this dilemma IMHO. The value of the overall network grows with the user base, so adding new federation partners should be near-free. Each instance is small relative to the network as a whole, and thus can focus on individual user value rather than growing the network, which means it can charge users directly. This is why you’re willing to buy a great Twitter or Reddit client app, but would never pay for a Twitter or Reddit subscription. (Yes I know those are centralized services, but the model holds if you look at the business relationships).