Many organizations are unsure how to demonstrate an ROI from a data center optimization program
In today's economic reality, even government organizations can’t initiate an IT initiative without first justifying the time and expense.
That usually calls for a return on investment analysis. Data center optimization is no different. Federal, state or local agencies usually need to conduct an ROI analysis before they can even start.
There is no magic to demonstrating the ROI for data center optimization. Judith Hurwitz, president of IT consulting firm Hurwitz and Associates, based in Needham, Mass., advises organizations to start by asking a basic question: “What are their goals? If they can't answer that, then maybe they shouldn’t get involved in data center optimization at this time,” she says.
Typical goals include reducing server or storage sprawl, lowering energy costs and reducing management staffing — in effect, doing more with less. Each of those goals can be quantified and translated into financial terms.
For example, the federal government's data center consolidation program has clear quantitative goals. Federal CIO Steven VanRoekel said in November 2011 that the government hopes to shutter 962 facilities by 2015. Previously, the Office of Management and Budget set a goal of shuttering 373 data centers by the end of 2011 and 800 by 2015, with a goal of saving $3 billion annually. Agencies published their latest plans for consolidation in fall 2011.
The government has closed 81 data centers by the end of November 2011 and envisions saving $630 million in the near term with the current consolidation plans and $5 billion in savings in the years beyond 2015, VanRoekel said.
Although data center consolidation is a goal primarily of federal agencies, the financial benefit is not clear for many government officials. Roughly half of the respondents to a fall 2011 survey of 321 IT officials conducted by the 1105 Government Information Group say they need to do a better job of tracking important performance metrics, the precursors of determining a financial benefit to consolidation. However, organizations which had done feasibility studies in advance of their data center optimization programs were much more confident of the actual benefits.
Indeed, determining the ROI is a challenge. One-third of the respondents reported that their agencies were unsure of how to even demonstrate an ROI from a data center optimization program.
“It’s a situation that every organization throughout the federal government is going to deal with: quantifying how much we’re saving,” explains Donald Adcock, the Army Information Technology Agency's executive director in an interview with an 1105 Government Information Group publication in November 2011. “It’s really hard to say [because] the savings are so second- and third-order effect in the process. Because we have state-of-the-art computing facilities and the way air flow works, we drive down HVAC capacity and use of electric. Then you have to go back and say, ‘Well, how did we track it before? How much did we pay for electric?’ In a building like the Pentagon that’s complex with multiple tenants that control those things, it’s very hard to pull that thread and come up with how much we actually saved in a dollar amount.”
The indirect costs, such as energy cost savings, are much more difficult to determine than the direct cost of fewer servers, software licenses, disk drives, and other hardware and software. Data center optimization, by definition, should deliver an ROI, and demonstrating it should be a no-brainer. All it takes is to identify and track some easily monitored processes. “You want to reduce the number of servers, increase server and storage utilization, and reduce the number of administrators,” says David Kelly, president of Upside Research, based in Newton, Mass. This shouldn’t be that hard: tally up those numbers and put a value on it, and you have the beginning of your ROI assessment, he adds.
For example, by reducing the number of physical devices in the data center, server and storage virtualization can reduce floor space requirements, saving on rent and reducing energy costs. However, many agencies often aren't able to spotlight the ROI gains even when they achieve them. Nearly half concede that they need to improve how they track crucial data center performance metrics, such as metrics relating to utilization efficiency or power consumption. Experts say utilization numbers can be readily gleaned from management tools or even system logs.
The Army IT Agency could serve as a model for successfully trying ROI to the direct costs of data center consolidation, at least according to its executive director. It has eliminated more than 30,000 square feet of data center flooring, reduced the cost of software licenses by 10 percent, increased processor performance by 40 percent and increased server virtualization capacity by 30 percent, Adcock says.
In addition, one group of government data center managers who responded to the survey reported having an easier time determining data center optimization ROI. Those government agency officials who have had feasibility studies conducted by independent third parties for topics such as server and storage virtualization, deduplication, and power/cooling — 46 percent — report that they can demonstrate ROI as a result of their optimization efforts. Naturally, they also are more concerned with using critical data center performance metrics to gauge the effectiveness of their efforts. Similarly, those who have conducted a feasibility study are twice as likely to disagree with respondents who complain that the ROI from data center optimization is difficult to demonstrate (46 percent vs. 26 percent).
Returns beyond cost reductions
A good data center optimization ROI results from more than just reducing costs or increasing efficiency. Many organizations overlook how data center optimization can positively impact their operations. Improving a government process that increases customer satisfaction should be part of an ROI calculation. And data center optimization can provide IT managers and agency executives with the opportunity to quickly improve their operations by introducing new services in an inexpensive manner.
“There is another metric that is usually missed in the optimization discussion,” Hurwitz says. That is the return on opportunity. With an efficient, optimized, virtualized data center, it is easier to deploy virtual IT resources to support a new initiative, and it costs almost nothing to roll out such virtual resources. If the initiative succeeds, the agency wins; if it doesn’t, it is a simple matter to take down the virtual resources and use them for another initiative.
Thus a ROI calculation for data center optimization should include the value of an agency becoming more dynamic and agile, an organization that can introduce new services quickly. That is an admirable goal that should be part of the foundation for demonstrating ROI.