Cloud Adoption Curves

By Guy Currier  |  Posted 05-05-2011

Cloud Spending: The Hidden Truth

Cloud computing isn't a technology, or even a set of technologies. It is a newly understood paradigm for approaching application delivery. This is why every vendor you can think of has been saying for the past two years that it has cloud products: It does. The timeshares of the 1970s were cloud products. We just didn't put them in the wider context of the global communications network. Now we do.

This is a pretty useful insight when it comes to assessing how we budget and spend on cloud solutions, the topic of a study fielded by CIO Insight in March and April 2011. The survey received responses from 270 individuals involved in spending on both public and private cloud computing in organizations with at least 50 employees; 95 in organizations with 50 to 249 employees; 85 in organizations with 250 to 2,499; and 90 in organizations with 2,500 or more. Most, but not all, respondents had IT titles (82 percent). To view our complete report, "The Hidden Truth About Cloud Spending," including charts and graphics, click here.

Recognizing that cloud computing is really a framework for understanding application delivery helps us interpret and adjust our strategies. We can look at where the spending is taking place--not in terms of use case or business unit, but by budget area. As with our survey on security spending ("Enterprise Security Spending Trends," CIO Insight, January/February 2011), much of our focus in this report is on "hidden" spending--that is., spending drawn from areas other than a dedicated, or even centralized, budget.

Taking cloud computing as a framework, not a technology, helps even when you consider private cloud computing, for which only 42 percent of our survey respondents have dedicated budgets, compared with the 47 percent of respondents that have dedicated public cloud computing budgets. (See Finding 1.1)

Partly, this is just indicative of where we are in the private cloud adoption curve.

More so than for public clouds, private cloud deployments are being driven at the application level--one of the more interesting aspects of the current cloud computing market. Individual departments or lines of business that most need the private option are the ones allocating the spending, and for one particular use. (For public clouds, many applications, even when departmentally driven, can be easily offered as a service to the entire organization, creating opportunity for centralized spending.)

Cloud Adoption Curves

So, what will happen to all these individual private clouds as we move along the adoption curve? We do anticipate ongoing centralization. For example, today's ERP cloud will become tomorrow's ERP-human resource management cloud, and the day after tomorrow's corporate private cloud for general company use. The biggest share of dedicated private cloud budgets today comes from hardware investment. (See Finding 1.2.) Nearly all organizations have hardware as part of their private cloud budget. (See Finding 1.3.)

In private cloud environments, average spending per employee on necessary hardware is $871, much higher than the average $510 they need to spend per-employee on IT staff to operate their private cloud. (See Finding 1.4.) It's important to keep in mind, though, that these systems, network devices, appliances and other needed private cloud equipment shouldn't have to be deployed anew for every new private cloud application. So eventually, as usage goes up, the overall costs are likely to ease.

The spending scenario is far different in public cloud spending. Hardware investment--in the form of upgraded Internet services and new client systems--is a smaller share of budget, accounting for an average of $597 per employee. But that's not the only factor to consider. There's an extremely low barrier to adding new public cloud applications when you need them, independently of any others. And there's no particular economy to be found by adding an application to an existing public cloud setup.

This easy proliferation of public cloud applications explains the relatively high level of staffing costs as part of public cloud budgets (an average of $792 per employee).

The result? Organization and management become critical issues. And this need for staffing is an ongoing cost, not something you can write off after startup. This is a further reason to expect public cloud budgets to remain stable and not decline at the levels of their private cloud counterparts.

Much of cloud spending is drawn--and will continue to be drawn--out of non-cloud budgets. This is a perfect example of how cloud computing functions as a framework for building corporate service strategies, rather than serving as a technology unto itself.

Budget Areas to Watch

The most-affected budgets--ones in which half or more of respondents have reported spending increases because of cloud computing needs--are relatively straightforward. Here is the percentage of respondents who say they are spending more, due to cloud computing needs, out of these four key budget areas (See Finding 2.1.):
  • virtualization (55 percent);
  • IT security (51 percent);
  • compliance/governance (50 percent); and
  • networking equipment (50 percent).

Virtualization, networking and security are obvious areas that need to be buttressed to support effective cloud computing. What's interesting is that for the first and last of the above budget areas, which serve as foundations to a cloud computing infrastructure, a moderate number of respondents also report that their spending needs had decreased--no doubt as a result of new efficiencies being found with holistic approaches to distributed computing.

Compliance and Governance

Though third on the list above, we consider compliance and governance budgets to be affected the most by cloud spending out of all budget areas analyzed in this survey. Why? Because the compliance budget is the area that is least likely to see any decreases in cloud spending.

Another similar spending area is mobile devices: 47 percent of respondents say cloud-related spending increases this budget category. And, when asked how actual 2010 spending was different--either higher or lower--than what was budgeted, respondents reported that, as a result of their organization's cloud deployments, they spent 8 percent more, on average, than was budgeted for mobile devices directly. (See Finding 2.2.) For that reason, we think mobility is a key area to watch as you plan your cloud spending in the months ahead.

Another area to watch is the applications development budget. Here, cloud initiatives are bringing 2011 spending 6 percent over what would otherwise have been budgeted in this category. This area is undergoing a transformation to service-oriented models, as well as to mobile-oriented models, as a result of the proliferation of cloud infrastructures.

While 44 percent of respondents say that their server hardware investments need to increase, nearly a third of respondents (32 percent) report that these investments have actually been decreasing. This means that servers are a likely area for rationalizing in organizations where existing installations are sufficiently up-to-date.

We suggested above that the growth of private cloud computing budgets should be tempered as the adoption curve progresses. We're not at that part of the curve yet: Our survey shows robust growth in spending of 15 percent in the typical organization in 2011 compared with 2010 levels. (See Finding 3.1.)

Will this significant startup investment be mitigated at all by so-called "hybrid" clouds? First, a little perspective on hybrid cloud computing.

We've noted that cloud computing isn't a technology; frankly speaking, hybrid clouds are not really a type separate from public or private clouds. Anywhere the latter two are used and show any interaction--simple enough with the extensive development of pre-cloud Web services, service-oriented architecture, CORBA or any other middleware--is an example of a hybrid cloud.

Roughly half of respondents (49 percent) have dedicated spending on at least one application that combines public and private clouds. (See Finding 3.2.)

It's important to note, though, that more than half of respondents (52 percent) say that this hybrid approach will have no impact on spending. (See Finding 3.3.) When it does, it's most often to reduce costs and usually on the private cloud computing side. In other words, this is smart use of clouds: private when you have to (which is pretty often) and public otherwise, because public is a lower-weight investment. You can apply this on a feature-by-feature basis if you wish, and not just wholesale for entire applications.

Where Do We Go from Here?

What are the key takeaways as you plan your cloud spending for 2011 and beyond? It's important to bear in mind all the enterprise budget segments that are affected by the cloud--and involve the appropriate business-side stakeholders early and often. In addition, keep your eye on these key budget areas: virtualization, security, compliance, networking and mobile devices.

Beyond this, our findings show clearly how cloud computing is a paradigm that can permeate all areas of technology activity. So when you consider all the spending outside of dedicated cloud computing budgets, you might realize that centralized planning can go only so far before it starts to hinder development. All areas of the business can absorb, and react to, the new paradigm. So the "hidden" truth about cloud spending might actually be that hidden spending should be encouraged, not avoided.

For more information, including charts and graphics, download the full report, "The Hidden Truth About Cloud Spending."