Gauging IT investments
A new process to evaluate information technology investments using science, statistics and mathematical models is getting so much attention that even the government's statistical experts will take it for a spin early next year
A new process to evaluate information technology investments using science, statistics and mathematical models is getting so much attention that even the government's statistical experts will take it for a spin early next year.
Census Bureau officials plan to award a contract to Douglas Hubbard, founder of Hubbard Decision Research, for a pilot project evaluating data-storage networks using Applied Information Economics (AIE), a rigorous new analysis tool that he designed.
Census is the third federal agency to try Hubbard's process. Agencies are under pressure from the Bush administration to justify their investments, and AIE could help them avoid the costly mistakes of the past, according to a number of government officials considering the technique.
Previously, agencies just guessed how effective an IT project would be, sometimes investing in one that fell flat. "There's no doubt we need better tools," said Duane Donnell, head of the IT planning staff for Census' Information Systems Support and Review Office.
"Now that we're being forced to look at IT costs and provide a return on investment analysis...we are looking at methodologies, and this might be one of those," he said.
In this age of tight budgets and scrutinized spending, AIE can help agencies evaluate large and risky IT investments before they are built, as well as comply with the Government Performance and Results Act of 1993, a law designed to bring performance measurement to government, Hubbard said.
Much as an investment adviser looks at a stock portfolio, Hubbard uses AIE to analyze whether a project is cost-effective. "It is a very specific evaluation," he said. "It is a balance score, and it is probably the most rigorous method for evaluating an IT method."
AIE determines what to measure when evaluating a project and assigns a value to each variable within the proj.ect, such as the number of known viruses per year that could hit a security system or the cost of each investment, such as firewalls or encryption.
Census could use this kind of analysis to decide what kind of data-storage facility to buy, how much storage it needs, whether to make the storage facility available around the clock or how to expand the size when needed. The agency plans to award Hubbard a contract of less than $100,000 to conduct the pilot project evaluating data-storage networks.
Private industry has discovered the methodology, too. The Discovery Channel and American Express Co. are two of the firms using AIE, Hubbard said.
Other federal agencies have already put AIE to the test. The federal CIO Council tried it out in evaluating an information security project at the Department of Veterans Affairs from November 2000 through March of this year. After Hubbard looked at the project plans, he recommended cutting a redundant intrusion-detection system, saving the VA $30 million.
Hubbard also examined a Federal Technology Service procurement project in July and suggested that the General Services Administration continue negotiating with a vendor to develop a Web portal and then re-evaluate the project.
The process offers government "useful performance measures and provides lessons learned for other federal agencies" to use, the CIO Council said in a September report about AIE.
Performance measures are important to the success of IT projects, said Lester Diamond, assistant director in the information management issues group at the General Accounting Office.
"The government has demonstrated a lack of good IT performance measures up to now. And there has been a surge of interest from the Office of Management and Budget and the CIO Council," Diamond said.
Unlike some methods that produce arbitrary scores, AIE uses real measures based on proven methods. At the end of the analysis, the IT project gets a score.
"We might say that you have a 92 percent chance of exceeding the cost of capital, or there is a 3 percent chance of getting a negative return," Hubbard said.
The process also takes into account such intangibles as policy-based budget constraints. For example, Census cannot afford to waste money even though it needs a centralized data-storage network that does not reside on disk technology.
"What is unique about AIE is that it factors in real-world things," said Patrick Plunkett, a program analyst with GSA's Office of Governmentwide Policy and co-chairman of the CIO Council's subcommittee on IT performance management. "You don't know the costs or the benefits, but AIE methodology allows you to evaluate that uncertainty."
***
Measuring the value
The Census Bureau is the latest federal agency to try a tool called Applied Information Economics to evaluate the cost-effectiveness of information technology investments. AIE relies on science, statistics and mathematical models. The concepts behind it include:
* Everything is measurable — even uncertainty and risk.
* Using range estimates for costs and benefits to estimate the value of IT is better than using averages or best guesses as estimates.
* The value of information needed to make decisions can be computed.
* Scientific methods of measurement are practical for IT investment.
NEXT STORY: Forman promotes e-gov resource sharing