One of the biggest selling points of cloud is its potential as a cost-saving technology. But where some CIOs err is assuming an easy correlation between cloud use and lower costs. James Staten, vice president and principal analyst with Forrester Research Inc. in Cambridge, Mass., sees this happening in companies of all sizes. Here, Staten shares with SearchCIO.com four questions CIOs need to consider when seeking cloud cost savings. Agility, existing legacy systems and current staff all factor into the calculus behind cloud ROI.
Do I have a firm grasp of how often the cloud service will be used?
James Staten: If the cloud service is going to be used in a perpetual fashion -- that is, if it's always going to be turned on, if every employee in the company will be using it and the amount of usage will rarely fluctuate -- it is unlikely that money will be saved. [Enterprise resource planning] would be a classic example, because traditional ERP systems have a fixed infrastructure footprint and can't take advantage of the elastic nature of a pay-per-use/pay-per-hour platform.
You have to look at what additional capabilities, functions, services and skill sets of consultants, exist between the two services, because that's really where you're going to get a cost savings.
James Staten, principal analyst, Forrester Research Inc.
Savings come with really understanding the cases in which it makes sense to use a cloud service in a "less than regular" fashion, with a few people in the company and where there's high fluctuation in the application's use of resources. Some types of BI [business intelligence] would be good fits. For example, BI that is occasional and BI done in batch are great examples. What doesn't fit, in most cases, is traditional data warehousing, as this requires a persistent presence of the large data warehouse. Other good examples are websites, mobile back ends, social media apps, and any test and development work.
Have I figured out the cost of replacing an in-house system versus moving that system or service to the cloud?
Staten: In a lot of cases, you have to determine whether you'll save money by migrating or replacing something you've already deployed. For example, you have the Oracle E-Business Suite for your e-commerce, but are considering doing e-commerce either through a pure SaaS [software as a service] application, such as Salesforce.com, or migrating this Oracle architecture up to an Infrastructure as a Service. In those cases, it's natural to ask what are the ongoing or soft costs of managing the work as it is right now, compared with what those costs would be running on Infrastructure as a Service, or what your sunk cost is if you swap it out with Salesforce?
The challenge here is that while it sounds like apples to apples, it never is. Take the migration to a public cloud example. When you migrate to the public cloud, you probably don't need the same configuration of the application. The licensing will be different because -- particularly in the case of Oracle with the way they license for a per-hour platform -- that cost is absolutely going to go up. You're initially going to save a lot of money by not having to manage and maintain servers: Amazon is going to take care of that. However, you can't put a zero in the IT ops column for your organization that had managed that infrastructure. There are still going to be tasks you have to do. Doing a rough estimate, you could take it down by, say, 70%, but it doesn't go down to zero.
On the application management side, your team is probably going to manage the application the exact same way, but the configuration of the application -- how they do performance management, capacity management and scale -- is going to be different. In a lot of cases, it will be cheaper, but in some cases it won't. It depends on the complexity of the application, what other systems it talks to and so on.
Have I factored in all the soft costs of cloud computing?
Staten: There are a lot more soft costs to be looked at than there are hard costs in the cloud, because it's a service. Let's say I'm looking at GoGrid and Amazon for Infrastructure as a Service. I want to do an apples-to-apples comparison so I'll be able to tell which one will save money, but once again the services are not actually apples to apples. You have to look at what additional capabilities, functions, services and skill sets of consultants exist between the two services, because that's really where you're going to get a cost savings.
More about how CIOs are cutting costs
Where CIOs see best cost-cutting opportunities in 2013
Vendor contract management keeps costs in check
Why CFOs love cloud metrics and when CIOs need to push back
The push is on for cost transparency in IT
Factoring in your in-house expertise in cloud adds another wrinkle. Let's say I want to get started in the cloud with one app, and I've got two developers in-house who can probably handle this project pretty well: That might be an apples-to-apples comparison. But later, you may want to add five more apps or you may want to do a more complex application and need help from third parties. If you can find a developer who knows Amazon and can be productive on your platform right away, the calculation is pretty straightforward, but if it takes six months or a year to find someone who knows GoGrid, then which one saves you money? These are significant factors that need to be weighed to truly understand your costs.
Have I factored speed-to-market into my calculus for using the cloud?
Staten: Agility is also a cost factor. If the purpose is to deliver mobile apps for the company, and it's done in the cloud using a native cloud mobile application or service that supports mobile apps, it's probably going to get to market within a week. That is a huge cost savings, and the company is probably going to make more money faster by doing that. This is probably more important than analyzing whether to do something in the cloud or on premises and figuring out the long-range cost.
A lot of people still want to do three- to five-year costing of things -- typical ROI calculations -- but very rarely does a company stay static for three to five years. Take customer experience, for example. Five years ago, if you made a decision about being on the Internet, you probably would've assumed that you'd need a website for your company and that's it. Now, in this fifth year, you probably have multiple websites with tons of pages behind them, you take orders directly off the website, have a collaboration site online, have support online, and probably between four and ten SaaS applications in use. So what was the ROI on that initial investment? It's kind of hard to know that. That's very much the reason you need to think more about what your company will gain from a cloud investment right now. And if you're going to look long term, look only about two years ahead.
Let us know what you think about the story; email Karen Goulart, Features Writer.