I spoke on a webinar recently exploring cloud ROI, and below are six key conclusions I took away from the discussion. I had a different view from one other webinar participant on the last one, about talent strategy related to cloud computing.
Here are my six takeaways:
1. The cloud has arrived.
Obviously, the cloud is no longer a fringe use case. Large, mature organizations are adopting it.
2. Discernment is needed.
At the same time, the cloud isn't for every use case. Pushing cloud computing on your organization without a clear business case or benefit is insane.
3. Brokering is vital.
Because discernment is needed, IT leaders need to foster a broker mentality. Someone in the IT organization must be an expert in cloud software and provider capabilities and connect use cases to the best services. Some workloads are inherently better suited to a public or private cloud (or, for that matter, to a particular flavor of public cloud).
[Feel like you're juggling chain saws while walking a tightrope? Don't make these 3 Common & Costly CIO Mistakes.]
4. ROI leadership: Influence, don't rule.
Only 23% of IT managers in a 2014 InformationWeek cloud ROI survey say that they can accurately compare costs of on-premises versus public cloud operations. And 51% say that they do compare those costs, but it's only a ballpark estimate. How can business leaders rely on such rough estimates?
It's because the financial ROI isn't the only evaluation criterion for a project, and the right business case doesn't need much of an ROI calculation.
My InformationWeek colleague on the webinar, Chris Murphy, quoted Dave Augustine, the Airbnb manager of site reliability engineering, on what he thought about calculating cloud ROI: "Right now, we have better things to do." I've heard that sentiment from other startups: There's such a clear and compelling business case for new infrastructure, and such a rich collection of practices with startups that show that the cloud is a factor in success, that it would be unproductive to work up a complex, custom ROI study for a specific provider.
The way I see it, we do plenty of things that become best practices that don't require gigantic ROI studies. Does anybody do ROI studies on VoIP anymore? No. We're at the same tipping point with the cloud.
5. Business transformation trumps incremental ROI.
Organizations perform most ROI calculations using financially conservative principles: Never overstate revenue or understate costs. And don't factor in "soft" costs. This is good. But transformational efforts can get understated in these cases, which is bad. How can you put a price on business agility? It's really hard to do that, yet most executives would agree that speed to market is not just important, but rather a critical factor that could spell the difference between failure and success.
One reason cost-benefit spreadsheets haven't replaced the C-suite is that wickedly difficult judgment calls aren't driven by financials alone. They require humans (for now). When that judgment says that cloud computing will be transformational for reasons having to do with agility, pay-as-you-go, or geographical diversity, you sort of have to take the spreadsheet with a grain of salt and move on.
6. Clear cloud talent strategy is needed.
This one's a no-brainer. But I disagreed with the speaker from the webinar's sponsor, VMware, on one aspect of the talent imperative.
He made the case that using a virtualization product that has ties into a hybrid/public cloud environment makes the transition to cloud computing a lot easier from a skills standpoint.
It doesn't. I say this based on years of experience, having dug deeply into three different hypervisors, three different cloud stacks, and two different public clouds. I've found that product configuration skills represent only a fraction of the transformation skills needed to build cloud expertise. The bulk of the real change is in deployment, orchestration, thinking of servers as code, change management, and bringing a DevOps mentality to infrastructure.
Thinking of the cloud as a big bucket of hypervisors, where hypervisor management skills matter most, misses the point. Yes, you'll get some agility by not having to order and then rack and stack servers. But the thinking that goes into API-based scaling, new app architecture, and getting beyond a manual management console and into scripted and automated deployment -- that's the bulk of the cloud transformation.
IT leaders must focus on the universal ROI takeaways. But more than anything, make your decisions based on business needs, which include ROI but not exclusively ROI.
Who wins in cloud price wars? Short answer: not IT. Enterprises don't want bare-bones IaaS for the same reasons they don't buy many $299 PCs at Wal-Mart. Providers must focus on support, not undercutting rivals. Get the Who Wins In Cloud Price Wars? issue of InformationWeek Tech Digest today (free registration required).