Makes no sense. If you're using cloud-resources then you probably already have an IT staffer on the payroll.
The hard part is not detecting when you're over-/under-provisioned. The hard part is implementing the actual automation beyond "start/stop an instance".
Not when you don't have an IT guy. I mean, what is the scenario here? Some contractor left you alone with an elastic EC2 app, that scales magically simply by turning instances on and off? So you have this highly sophisticated application yet nobody who knows how to operate it, nor capacity monitoring? And your best bet is to seek some crummy outside monitoring to aid your... sales-guy(?)... as he pulls the levers?
Seems a little far fetched to me.
If you're a small startup with no tech clue then you're probably deployed on a shrink-wrapped solution (e.g. heroku) to begin with.
I guess I just don't see who they're aiming at here. Their marketing clearly suggests small business with no tech clue; but how did such a business scale to a point where savings are possible in first place then?
Still makes little sense. If you have no IT guy then you're likely not using big resources, simply because you wouldn't know how to set them up.
It's really, really complicated. That's partly because I don't think Amazon is particularly good at UI, but largely because AWS lets you do so much. They have taken things that have traditionally been sold together and split them all apart (IP addresses, compute, storage). Heck, they even sell elastic network interfaces now! Also, AWS was largely designed to be controlled through APIs (and that's how the heavyweight customers use it), so there's not much incentive to make the UI good.
But most customers don't need that level of control, yet they're using AWS to run their apps anyways, probably because they have the most market recognition. Now they have to deal with the complexity but without reaping the benefits, which is rather unpleasant. Hence the need for additional services to help them manage it all.
As the CTO of OpDemand, I can tell you first-hand that users are willing to pay for a simplified and streamlined AWS management process. Our users want a Heroku-like experience, but they don't want to give up control of their underlying infrastructure.
With that in mind, we focus on tight GitHub integration, a sleek UI (backbone.js/socket.io) and providing a toolbar with 5 dead-simple actions: start, stop, deploy, clone and destroy. We charge for our tool because we save AWS users time and money.
Sounds like these guys are building a little robot that watches the gauges for you and decides when to switch things on and off.