I’m not sure why I find this topic so interesting? I’ve been thinking on it, on and of for days.
The assumption that newspapers deliberately put in soft paywalls - would mean that there is an easy and clear way to model the positive outcome of such soft restrictions. The Lagrangen Multipliers are popular in economics and for what I’ve read even used by mafia to model optimum sums for bribery. i.e. how you would relax a system optimal by introducing another variable. I did try to find material how such a model would have been set up, but so far I haven’t found anything. Even tutorials on Lagrange Multipliers are hopelessly mirky.
I did find one example of a different kind of paywall that seems very interesting. Instead of paying access to content, you would buy access to comments and a community. So all articles and content is open, but you have to pay to read and take part in any discussion of any content. In this case, I assume there would be nothing to gain with soft restrictions? Instead, soft gain to content would probably upset members.