Pender's Blog

Blog archive

Google's SaaS Scare

Nobody's perfect, right? Not even Google, which has suffered some well-publicized service outages recently and finally felt compelled to apologize for at least one of them.

Hey, Google, love means never having to say you're sorry. And we still love you -- don't we? After all, you're the revolution, the flag bearer for Software-as-a-Service, the company that's going to have us using nothing but free-ish applications in our browser before too long and chucking that clunky old Microsoft software in the same bin where Corel WordPerfect now resides. Right?

Well, maybe not. It might not be a big deal to most of us -- and your editor does have a personal Gmail account -- that Gmail went down for a couple of hours on a weekday afternoon, but it was probably a pretty big deal for people who are trying to run their businesses on Gmail and Google Apps.

Beyond that, and more to the point for the channel and users alike, the mighty Google's stumble was an unpleasant reminder that SaaS as a model still isn't as ready for prime time as many folks would like it to be...despite all the SaaS hype that some media organizations (including this one) have spread recently.

OK, so Gmail going out for a couple hours isn't the end of the world for most companies, although e-mail outages can be a major problem. But what about enterprise applications? What happens when CRM goes kaput for a couple of hours on a Monday? What about -- heaven forbid -- ERP applications?

It's no wonder that hosted ERP is the last frontier for the SaaS model. Really, really critical stuff -- like the software that runs companies -- just can't be subject to significant downtime. Oh, sure, on-premises stuff has its problems -- huge, costly problems in some cases -- but at least companies have staff members or helpful Microsoft partners who can take care of issues in-house. When somebody else's datacenter goes down, what do you do? Sit and wait for somebody far away to fix it.

The real crux of this story is that we're talking about Google here. If Google, which is betting much of its business model on SaaS and pouring millions of dollars into datacenters, can suffer fairly nerve-jangling outages, what about smaller hosting companies with fewer datacenters? What about other vendors that don't specialize in hosting but are just dipping their toes into the SaaS waters (we're looking at you, Microsoft)? This sort of thing doesn't do much for users' confidence.

Partners need to have some sort of SaaS story for their customers. But they also need to know how to answer the inevitable questions about the model's Achilles heel. And they need those answers to be specific and meaningful, not just empty guarantees or platitudes about point-nines of uptime. It's one thing for customers to rationally know that SaaS comes with the risk of downtime; it's quite another for them to actually experience an outage.

It's something to think about now that the leader of the SaaS revolution has racked up a little bad publicity for itself and the model and turned a little bit of SaaS love into apprehension. After all, nobody's perfect -- but it's dealing with imperfection that makes some companies great and causes others to fizzle. Google is big enough and experienced enough to be great in the long-run. Not everybody else, though, will end up in that category.

What's your take on SaaS and the risk of downtime? Do incidents like the Gmail outage make you think twice about moving key enterprise apps to a hosted model? Have your say at [email protected].

Posted by Lee Pender on August 14, 2008


Featured