Ask some CTO’s about how their product scales and they’ll whip out a logical diagram showing you redundant networks, redundant firewalls, load balancers, clustered application servers, redundant databases, and SAN storage. If you’re lucky they’ll tell you about their software stack and then throw in a bit about their software development life cycle.If this is how your SaaS vendor responds to scalability without going into some operational specifics, then beware.

If you are a very early adopter of a SaaS vendor, then you’ll want to look at infrastructure. You’ll definitely want to be evaluating their leadership and their business plan.

But let’s assume for a second that you’re not one of the first ten customers. Scalability for a SaaS vendor has at least the following areas:

• Software platform – How often do they put out new releases of core software? How is it rolled out to customers – in one shot or in waves? How do they handle the QA of these rollouts – and specifically, do you need to participate in this testing.
• How do they monitor their systems for service level and capacity issues?
• How do they manage network operations? How is staffed? How fast do they respond and recover from issues? Can you see their issue log?
• How are customer specific implementations managed in their software repository? Beware of the runoffs!
• How are new features prioritized? How do they involve customers in the prioritization?

The point is, scaling is really about scaling for more customers and all service level dimensions, not just infrastructure.