Insight by Grant Thornton

A path to deal with data chaos and avoid mission failure

Aurpon Bhattacharya, a principal and Foundational Analytics Lead at Grant Thornton, said agencies can reduce the overwhelming feeling of managing and understanding the troves of data they own.

For much of the past decade, agencies and industry have talked about how data and the analysis of such data must run the government’s mission.

But even nearly 15 years after this initial recognition of how important data is, agencies continue to struggle with some basics, including understanding the data they have, what business questions do the data address or can address, and even if those questions are the right ones?

Aurpon Bhattacharya, a principal and Foundational Analytics Lead at Grant Thornton, said there are proven ways that agencies can reduce the overwhelming feeling of managing and understanding the troves of data they own.

Bhattacharya said agencies, and really all organizations, are experiencing a “data chaos,” and must figure out the path out of that disorder.

“Leaders and managers are overwhelmed with the amount of information they are seeing, the amount of data they are seeing. How do they manage that data? What kinds of insights can they get from the data that they are seeing?” Bhattacharya said on the program, Moving from Data Chaos to Data Insights sponsored by Grant Thornton. “One of the major things that needs to happen is to stop looking at the shiniest object that they see in front of them. Today we see Bitcoin, artificial intelligence, robotics process automation or machine learning are all a big deal. Do we jump at each one of them and think they are a solution? They are not a solution. They are a tool by which we can leverage data to address the burning challenges agencies are facing.”

He said one example of this challenge is the Veterans Affairs Department’s goal to reduce veteran homelessness. The objective may seem straightforward, but VA must first understand if they have the data on the back end to solve this problem by answering the mission-related questions.

“There are a lot of business questions that need to be answered so you have to interview agency leaders to figure out what is the hot topic that pushed their agency forward. Then you look at what data is available. It’s quite possible that the data they need to answer those questions isn’t available,” he said.

Agencies should take a three-step process to answer all of these questions:

  • Discovery–Finding the data that exists
  • Diagnostic—Understanding the data and figuring out where the gaps exist
  • Prescriptive—What does the data need to do so they can plan for the next 3-5 years

Bhattacharya said Grant Thornton used this three-step process recently for a large federal agency which wanted more insights into their purchase card data.

“We asked them what was the business question they were trying to solve? We figured out it was trying to prevent waste, fraud and abuse,” he said. “We used our repeatable fraud analytics solution and within a month had insights he never had before. What he found out was numerous purchase cards were being used over the weekend in a parking garage in New York City, over Christmas Eve, over New Year’s Eve and during holidays in general.”

Bhattacharya said previous analyses went line-by-line of the financial system data and never put together the trends that emerged by applying advanced analytics to solving a specific business question. Now that agency has a better idea of whether a fraud, waste, or abuse problem exists or if there are legitimate business reasons for these and other charges.

“There is a whole culture mindset to make analytics successful,” Bhattacharya said. “We need to look at where is the culture change happening—top down or bottom up? Who is really interested in driving insight using data? We are seeing over time that this culture of analytics is slowly taking place. It’s not just analytics, but a change management exercise in and of itself.”

Bhattacharya said the biggest mistake agencies and other organizations tend to make is not knowing the right questions to ask and focusing instead on the technology to answer the incomplete or off-center questions.