Software-as-a-Service is so common it's actually boring at this point, but there are still only a few very large SaaS companies. As SaaS companies grow, both in market and technical scale, the need to refine and mature processes becomes more important.
We really have no idea how the Operations are done in the Cloud, with the exception that we have a bit of visibility into the uptime/downtime. We still lack even the most basic release management visibility and we have no idea how Cloud providers deal with service interruptions, or how they do upgrades.
For example, there were recent issues with VMware ESX 3.5 Update 2, in enterprise configurations where the hypervisor wouldn't power on after being turned off. What would have happened if an entire Cloud system was shut down at that moment? Would customers even have been aware?
One of the big issues that I wonder about is just how well versed the operations teams are as the systems need to scale up with a very high SLA level.