Bringing Cloud ROI Down To Earth 2

The calculation on whether to outsource a given IT function must be based on data loss risk, lock-in and availability, total cost, reasonable investment life spans--and consensus on when good enough is all you need.
Any CEO can look good during an 18-month stint with a successful company in a hot sector. The real test comes over the long haul and during market fluctuations. That same extended perspective can be a sticking point for most return on investment and total cost of ownership calculations around newer technologies, including public cloud services. For example, those skeptical of the public cloud aren't convinced that swapping capital investments for ongoing operating costs will benefit IT long term. How much of an operating cost? How long of an operating period? Is it the same time span that you'd use to evaluate a capital investment, or is this a shell game, where cloud is less expensive in the short run but more costly over time?

It's not just an academic exercise. As any expert on corporate strategy will tell you, fiscal responsibility is about much more than short-term gains.

InformationWeek Analytics' April survey on cloud ROI takes a look at how nearly 400 business technology professionals view the financial impact of public cloud services. The good news is that IT isn't running blindly to this new model.

Cloud computing works for commodity applications, but any integration or complex configuration makes the cost skyrocket, says Jack Garhart, an IT manager with a large international financial institution. "Most of the companies I've worked with like their investments to last five-plus years. It seems that the breakeven point for investments for internal versus SaaS is three to four years. We end up paying less in the beginning, but then pay a lot more in the out years," Garhart says.

Respondents are clearly doing some ROI calculations. But the evaluation periods they're using don't always reflect what we'd call a business system lifetime, even if the factors they're considering are reasonable. So why aren't companies doing the the exhaustive discounted cash-flow analysis needed to study all the variables involved when switching from on-premises systems to cloud?

Mainly, because they're evaluating a moving target. While the National Institute of Standards and Technology did everyone a favor by creating umbrella definitions of cloud computing, when we start diving into details, there's little universal agreement on many areas. What exactly is a private cloud? Can software as a service ever run on one, or does that defeat the purpose?

You get the picture.

We're normally big fans of details. It's granular attention that keeps enterprise systems running. But in this case, we can't fault CIOs for not insisting on the same level of analysis they'd use for an established technology. This is not, however, a green light to give up on fiscal discipline. Instead, one trend emerges after studying our research and speaking with a variety of companies: There's a new breed of smart and disciplined adopter who loosely studies ROI and then revisits the calculation over time to ensure that expected returns are panning out.

"Customers are mostly in test and piloting mode: 'We understand the fundamentals; let's do a pilot to see if it's true and how much it's worth,'" says Marten Mickos, CEO of private cloud systems vendor Eucalyptus. The true discounted cash flow and net present value calculations will happen no sooner than a year from now, Mickos says.

Near term, figure out what factors are important for your company, calculate the specific returns for the implementation you're using, and keep monitoring to make sure reality is in line with expectations.

Risk And Reward

Our InformationWeek Analytics report on cloud risks looks at the biggest barriers to cloud computing adoption; fear of unauthorized access to customer or proprietary information, lack of maturity, and security defects top the list. But if risk can be managed, respondents see good reasons to use these services.

As you'd expect, speed of deployment is near the top of the list of perceived benefits. Lower long-term expenses also sound good, but if you're one of those folks who's building a case for cloud around short-term capital savings, do your homework. Two years of ongoing cloud expenses may be less costly than a one-time build of hardware, but is the comparison going to be favorable for the entire life of the system, perhaps upward of five years?

It may well be, but run the numbers, taking the system lifetime into account. Our full cloud ROI report, free for a limited time at analytics/cloudroi, includes a sample ROI worksheet to help with that.

Twenty-eight percent of respondents say there's inherent good in replacing capital expenses with operational expenses. That reminds us of the old PC leasing argument: "If we make PCs an operational expense, then we can't get denied capital--we're in a contract, and we can't get out of it."

That's true enough, but today many companies no longer lease, and not just because they're looking to avoid paying interest. They've also brought their CFOs into the conversation and know that a periodic capital expense can be less costly than an ongoing operational expense. Our point is, capital expenses aren't inherently evil, as long as credit or capital is available, and operational expenses aren't inherently good. It's all about the present value--that is, the normalized value of the future cash outflows in today's dollars, taking into account the potential cost of borrowing.

A small but crafty group of respondents, 14%, would like to move computing expenses off IT's budget and onto line-of-business balance sheets. There's a lot to be said for this idea. The fact is, expressing any purely line-of-business expense in the IT budget misrepresents how much IT costs a company. Our take is that activities that are engaged in for the benefit of only one division should reside in that group's budget, because unless there's another way of encumbering that business unit--like a chargeback--representing these activities on IT's balance sheet hides the true cost of that division.

You don't need cloud computing to make this happen. But the cloud could provide a clean break. One caveat: Off-loading the expense and function to the cloud doesn't mean that IT will be held harmless if the business gets hosed--for example, by being locked in to escalating prices. In our experience, outsourced IT services that go sour always get blamed on internal IT, no matter who picked the provider.

One final note from our survey: We found it troubling that, among companies not using cloud computing, the CIO is the leading skeptic rather than the chief enabler and information gatherer. In fact, the CIO is almost twice as likely to be the naysayer than the next most significant "no" vote.

"CIO" may soon stand for "Cloud Implementation Opposition."

It's very likely that CIOs who are flat-out saying "no" are being bypassed by business units that do their own technology implementations. If it's not happening yet, it soon will be. That also might explain the gap we discovered when we asked business units versus central IT, "Is IT involved in cloud at your organization?" We found that 94% of IT pros said yes versus 79% of those in business units.

Don't say you weren't warned.