Insight by Gordian

How agencies can estimate construction costs more accurately

This content is provided by Gordian.

Construction cost estimating is a complex and multi-layered challenge that almost every agency is trying to get a handle on. They’ve been largely hindered in the past by a weak reporting framework and inaccurate or obsolete cost data, as well as consistent budgetary uncertainty. The problem is getting so serious that it’s gotten the attention of Congress and the Defense Department.

“Just take a look at our major threats around the world. China — building an island in 20 days. We can’t even get a contract written in 20, let alone get something built. So I would tell you this is a national security issue,” said Ret. Brig Gen. Joseph Schroedel, executive director of the Society of American Military Engineers during a Gordian and Government Executive webinar on federal project overruns.

One problem is that the agencies are attempting to predict the cost of construction anywhere from three to seven years in advance. But acquiring timely and accurate construction cost data to do this has always been challenging for federal agencies. Inflation, commodity price increases, and unexpected costs or delays can all cause projects to overrun, making it difficult to make informed decisions about construction costs.

But the situation isn’t hopeless; some agencies have discovered some solutions to these challenges.

“We’ve had a lot of disaster scenarios for our Federal Emergency Management Agency, or FEMA, and when hurricanes hit our tropical areas — places like the US Virgin Islands or Puerto Rico — how we do estimates about getting relief to those places, dealing with those things requires a lot more thinking and data and application and understanding what the work is, and using work breakdown structures to get to those more valid or real assumptions,” Tim Persons, chief scientist and managing director of the science technology assessment and analytics team at the Government Accountability Office, said during the webinar. “So data driven assumptions are key. And so there’s some success at FEMA having adopted some of these things where [they] never happened before.”

To reduce risk in their construction grants after Hurricanes Irma and Maria, FEMA contacted Gordian to help develop high-resolution cost data for Puerto Rico and the U.S. Virgin Islands to add to their RSMeans database, which FEMA relies on heavily for cost estimations.

Building on that and other best practices, the GAO has published a cost estimating guide for the federal government.

“We’ve got to gather data, in order to capture experience. Beyond capturing the data and capturing experience, we’ve got to be able to know how to apply that experience in the proper way,” Schroedel said. “In terms of data driven methodologies, I think it’s less about what data and which data elements, and let’s start with the outcome, we’ll start with the delivery part of this discussion. … How do you get the right data and integrate that data in a way that helps us manage risk?”

One way is to leverage technology. Cost estimators currently manually sift a database of more than 80,000 unit prices. Technology can help search through that database faster, finding specific line items while also suggest similar line items. And artificial intelligence has the potential to look at contextual costs.

In addition, skilled estimators in the federal government are nearing retirement along with much of the rest of the federal workforce. These technologies can supplement their experience in bringing up younger estimators through the ranks, who may have difficulty understanding how these projects go together. Junior estimators can use real language around what needs to be done on projects, and AI can assist them in finding line items.

But this also requires transparency. Currently, everything is siloed; every agency, and every region within those agencies are responsible for their own program estimates. Some have central repositories, but even those are generally antiquated technology, and won’t allow for analytics on all inputs. That’s one of several reasons FEMA relies on Gordian’s RSMeans Data Online for cost estimation data.

Improving transparency, however, requires a culture change throughout agencies, but more importantly starting with leadership.

“It really is about risk management at the end of the day. And I think leadership means you have to own some responsibility on things,” Persons said. “You have to really deal with the idea of the incentives or disincentives of saying, ‘here’s what it’s really going to cost.’ … It does require a different way of thinking. And it does require a risk management and leadership-based framework.”



Sign up for breaking news alerts