Could big-data analytics improve federal procurement?

Big data could be an important tool for federal procurement shops, but its usefulness depends on finding quality data and understanding how to use it to track vendor performance and pricing.

Several recent studies -- a Government Accountability Office report, a CIO survey by TechAmerica and the IBM Institute for Business Value's "Chief Procurement Officer Study" -- all point to the same conclusion: Analytics and acquisition need to go to more of the same parties.

The Oct. 9 GAO report states that many agencies' incomplete methods of performing market research affect their ability to make informed decisions about procurements.

Federal agencies are required by law to conduct market research, which the Federal Acquisition Regulation defines as the process used to collect and analyze data about capabilities in the market that could satisfy an agency's procurement needs.

However, auditors found that the agencies GAO evaluated -- the Federal Aviation Administration and the departments of Defense, Homeland Security and Transportation -- did not consistently perform market research and report on those findings.

"All four agencies require that market research be clearly documented since it can inform current and future procurements, but the guidance varies in terms of the elements that are to be included in the documentation," the report states.

Furthermore, big data could help procurement shops make the right decisions -- it they have access to it.

"Making data available will be key," said Doug Bourgeois, vice president of end-user computing at VMware and former executive director of the Interior Department's National Business Center. "In large departments, if datasets are merged, we will be able to track vendor performance, pricing and where cost savings can be achieved across the different bureaus and agencies."

Unlocking insights

In TechAmerica's 2014 federal CIO survey, respondents identified analytics as one of their top challenges and areas for future expansion. Although acquisition was also cited as an area in need of major improvement, analytics and acquisition were seldom mentioned in the same sentence.

The IBM Institute for Business Value's 2013 "Chief Procurement Officer Study" found a strong relationship between performance and the use of innovative technologies. Specifically, 94 percent of top-performing procurement organizations are highly effective in their use of technologies, which includes the ability to analyze big data.

"They are using big-data analytics to unlock insights to help reduce costs, drive compliance, mitigate risks, improve business intelligence, and manage and develop suppliers," said Matthew McGovern, market segment manager at IBM Procurement and Contract Management Solutions, in a post on MyPurchasingCenter.com.

"Procurement leaders are finding success in leveraging big data to accurately and more easily forecast demand, to identify unimagined areas for cost savings, and to identify suppliers that pose greater risks to the business," he added.

During Bourgeois' tenure in government, which included a stint as CIO at the U.S. Patent and Trademark Office, he said there were concerns about the accuracy of reports on the past performance of contractors. But by using big datasets, procurement shops should be able to run a more thorough analysis of vendor performance and continue tracking it, he said.

Prove the priority

Developing systems that can analyze big data does not have to stretch agency budgets, said Aubrey Vaughan, managing director for public sector at Oversight Systems.

"The U.S. government has been doing big data since the 1980s," Vaughan said. "Today agencies have the power to apply what we call out-of-the-box custom analytics that could quickly draw analysis right at their fingertips in minutes."

Instead of hiring consultants and building new capabilities, Vaughan said agencies should consider using firms that have already built analytics models. He added that Oversight Systems' software, which tracks vendor interactions, has saved the Defense Department $8 billion since 2008.

In addition, challenges and prizes could prove to be effective, Bourgeois said. For instance, agencies could use Challenge.gov as a type of modified procurement process by providing data and asking vendors to develop software to solve a problem. Whoever comes up with the best solution wins the contract.

Regardless of who gets the ball rolling on the big-data effort, Bourgeois said it should involve a broad policy group.

"It would have to be driven by governmentwide effort," he said, adding that the Office of Management and Budget and the General Services Administration "would need to team up and prove it's a priority for the administration."