Predictive analytics provides the likelihood of the future based on the past, and government agencies such as the Defense Logistics Agency (DLA) have recently begun to use this technology to optimize operations. Efforts don’t have to be large to see transformational effects, but they start with a good foundation in data.
Challenge: The Sheer Amount of Data
Data has become ubiquitous and prominent in people’s lives. Online stores recommend products based on purchase history; streaming services recommend shows after one unfortunate binge-watch. Whether the data history is good or bad, representative or unrepresentative of you, this is what predictive analytics does — makes predictions about the future based on the past.
The past, therefore, is critical to successful analytics. The quality of the data that an analytics project is based on — whether it’s descriptive, diagnostic or predictive — is the key and the challenge that agencies face when it comes to adopting analytics. Today, decisions are increasingly made with data in mind, from improving digital experiences for users to clearing the snow for residents to getting supplies to U.S. warfighters worldwide, which is DLA’s responsibility and mission.
“A challenge we had — and we still face — is just the sheer amount of data we have,” said Teresa Smith, Chief Data Officer (CDO) at DLA. The troves of data that the agency collects and the lack of structure to manage it made it difficult to determine whether the data was clean or authoritative enough to use in analyses for optimizing the supply-chain processes for warfighters.
Solution: Good Governance for Advanced Analytics
The CDO role has been pivotal for managing better data quality. The establishment of the office and position at DLA three years ago has encouraged a better data foundation by providing governance, the formal management of data as an asset. The structure has also allowed for more advanced analytics, such as predictive analytics, to take place.
Predictive analytics goes a step beyond showing why something happened by presenting what might happen in the future — or “what-if analytics,” Smith said. For DLA, it’s a helpful tool when it comes to filling orders for certain resources, including rations and equipment, for DoD personnel around the world.
“Like any organization, we have a finite budget, and so we want to make sure that we’re maximizing the return for the warfighter and providing them optimum readiness,” Smith said.
To do so, the agency built a material availability predictor model that shows a forecast of demand for materials up to a year in the future. It has helped the agency look at the holistic picture of its portfolio to glean insight for what actions best fit the future.
“We can look in the rearview mirror and see what [material availability] has been, what it is today and what it’s projected to be,” Smith said.
Outcome: Satisfied Employees
Currently, material availability metrics at the agency are at record-high. But it’s not entirely due to the technology. “It’s probably largely because of the hard work of the folks,” Smith said.
What predictive analytics has done, even in a small use case, is enable employees to anticipate problems and raise proactive solutions.
“Nobody likes to work constantly with their hair on fire,” Smith said.
This has brought more value to their jobs and led to more innovative solutions. Responding reactively often leads to resolving issues in the same way as before, which isn’t necessarily the best way. Predictive analytics allows more time for employees to develop creative solutions and get ahead of the problem. Essentially, if deployed well, the technology can add time, value and money to an operation.
“That’s the beauty of it,” Smith said. “We’re not going to catch every problem, but it does inform us ahead of time.”
This blog is an excerpt from our recent guide, “Technology Transformation Strategies: From Idea to Implementation.” Download the full guide here for best practices.