The pay-as-you-go model addresses several obvious and painful limitations of the previous model, which was based on asset purchase; in other words, prior to application deployment, a significant capital investment had to be made to purchase computing equipment (i.e., servers, switches, storage, and so on).
However, there is one big advantage of this approach: once the investment is made, the financial decision is over. Assuming the application obtains the necessary capital, no further financial commitment will be needed. Of course, this has led to utilization issues, as applications commonly only used single-digit percentages of the computing resource assigned to them, but there were no ongoing bills or invoices for the application resources.
[Related: ?]