> paywalls should federate, like the early ATMs did, so that joining one means joining them all
That’s an interesting idea. I imagine it could work sort of like Spotify Premium but with a twist. Perhaps that is not what he meant but it’s what I came to think of so that’s what I want to talk about.
Of course the real problem is getting all/enough/any of the paywall sites to switch. I’ll get back to this but for now let’s ignore that problem.
How the system will work: There is a payment processor (could be one of the existing payment processors or a new company) that provides the federated system to the paywalled sites. The payment processor handles billing of end-users and once per month it pays the paywall sites according to the proportional amount of articles (another problem I’ll get back to) that each user read on each site.
This is similar to the Patreon and other recurring payments services except that payment is calculated automatically similar to how Spotify does it instead of having the end-user decide the amount they want to give to each site. This is done for a couple of reasons. Firstly the point was to lower friction for end-users, so we don’t want them to have to individually set the proportions for each site. Secondly, if end-users were to set the proportions then they might forget to update them.
But there are more problems, both for the end-user and for the paywalled sites. End-users might derive more value from some sites than others. In the extreme case there could be two or more sites where one site had a really really good article that the end-user derived a lot of value from but in the same month they read a lot of content on the other site/sites that did not provide them with anywhere near the value of the single only article they read on the other site. And yet they end up giving most of their money to the sites that provided them with less value.
And that problem is further made bigger by this system encouraging the paywall sites to make more clickbait and to use other kinds of tactics to increase the number of articles read just as with the ad-supported model.
So one way to try and combat that is perhaps to use amount of time spent on each site instead of number of pages loaded. Amount of time spent would be a better metric if users simply navigated away as soon as they found some content to be of little interest. And we already know that they do so on the web in general. So it might work to a certain degree. At least it rewards the sites for content that users chose to stay consuming than content that they were simply tricked into loading but which they quickly determined to be not worth their time.
Of course it could still lead to articles that are longer and so take more time to read, while staying sufficiently interesting that readers stick to them. In fact I am pretty sure that this not just could but would happen. We’ve already seen sites split articles into multiple pages to increase the number of ad impressions they can serve. To instead insert filler content into interesting content within a single article is not much different from that. So while not optimizing simply for pageloads, publishers are still highly likely to optimize for time spent rather than as we ideally would want them to be optimizing purely for value to the end-user.
You could then have a scoring system that allows the user to score each piece of content, kind of like how HN and Reddit has upvotes and downvotes but where you use the scores that each individual user gave to each piece of content when you calculate the proportional division of money that will go from each individual user to the publishers of each piece of content that said individual user consumed that month. But first of all this increases friction once again, which is what we want to avoid, and secondly it encourages publishers to manipulate users into giving a high score. And publishers who made good content and did not remind users to “please give this article a good score if you enjoyed it” would once again get the short end of the stick. And on top of that nobody bothers to score every piece of content that they come by — but that can be accounted for with some maths, the main problem is the friction and incentivizing publishers to manipulate users about the scoring like I mentioned.
There’s a lot of other things to consider as well, but I’ll round off this comment now with my thoughts on adoption.
On getting sites to switch: I think at the core there are three factors that would determine whether paywalled sites would use this federated system.
One, revenue. Will the paywall sites earn more by using the federated system? With the lower friction of a federated paywall system, it could, but even if there are more subscribed users the sites are also getting only a portion of what each user is paying and the sum of those payments might add up to less than fewer users paying a site directly.
Two, adoption. In order for the system to work in terms of offering lower friction to the end user, high adoption among paywalled sites is required. In the beginning, before adoption is high, why would any paywall site agree to adopt it? A chicken and egg problem.
Three, trust. Will paywalled sites trust the provider of the federated system to be honest and pay them the amount that they should?