I think successful "pivots" like this usually have to do with rebalancing the risk of the customers. In the article's example, the risk is shifted away from the publishers ("am I sure this software will be worth it?") to the startup ("we only make money when this works").

To my eyes, even the more traditional enterprise-to-SaaS model shifts that are happening reflect the same trend - trading the (customer) risks of big $$$ up-front, long implementation etc for the (startup) risks of needing to attract more customers to make the same amount of $$$.