When Mark Ohlund accepted the position of CIO at Armada in 2011, he began leveraging the company’s available data to improve internal and external customer decision making. “To accomplish that goal, there were two areas in need of significant improvement,” Ohlund says. “The first was integration with our trading partners: distribution centers, suppliers, carriers, and customers. The second was the creation of a purpose-built data-warehouse environment to store and process incoming data.” Available storage space was quickly increased by more than twelve times through business continuity initiatives and the build-out of a data warehouse.
Improved data integration turned out to be a larger undertaking.
In logistics and supply chain management, most of the data is derived from inventory-related transactions: What is the current inventory level? Which inventory is at risk of expiring? Where is each particular item currently located within the overall supply chain? And most importantly, how do each of these factors relate to each other? When operating on a global scale, with thousands of daily supply chain transactions, a company can easily drown in a sea of data. To avoid this, Armada’s focus is on capturing actionable data. “If it doesn’t support part of a process somewhere within the organization, then it’s not actionable, and it’s not something we need to see,” Ohlund says.
Improvement in data accuracy following Armada’s implementation of the 24TRACC application
Increase in available storage space
Integrated trading partners using the 24TRACC application
In order to find that actionable data, Armada needed to get organized. In the past, the company had provided purpose-built scripts to individual trading partners in order to perform data integration. These custom-built tools were fragile, inflexible, and incapable of scaling. To solve this problem, the Armada team rewrote the company’s proprietary application, 24TRACC, following a software-as-a-service (SaaS) model. “Today, 24TRACC provides visualizations and reports while allowing visibility into customer inventory levels and supporting a number of supply chain related processes,” Ohlund says. For example, last-minute transportation costs can be avoided by identifying and avoiding their cause. “Just like airline bookings, the later you acquire that transportation, the more expensive it will be,” Ohlund says. By analyzing relevant data, Armada is able to observe patterns of behavior leading to increased premium freight charges and make recommendations that enable the company’s clients to acquire additional freight capacity at a reduced rate.
Armada also needed to address issues of data viscosity. “Historically in this industry, data gets transmitted in daily batches sent early in the morning,” Ohlund says. “We get one transmission of all the previous day’s activities, process it, and load it into our various applications for visibility.” By 9 a.m., that data may already be up to twenty-four hours old. “So our big push with our trading partners is to have more regular data transmission and eventually real-time data transmission triggered automatically by their enterprise applications,” Ohlund says.
In order to efficiently process various transmission levels, Armada deployed a new integration platform two years ago. “From a data-processing perspective, the biggest improvements are data-cleansing and redundancy-reducing activities, which make the data flow as pristine as possible,” Ohlund says. Preliminary results indicate as much as a 50 percent improvement in data accuracy over the earlier system. The automated data extraction also works independently of human interaction, increasing productivity and transmission speeds.
The new system marks an evolution in how Ohlund and his team approach data. “We are moving from a ‘rows-and-columns’ understanding of data to a visualization of data,” he says. Using actionable data to generate reports that display trends and inform customer decision making, 24TRACC processes the underlying forensic information. “I’m always asking the team, ‘What can this tell me in five seconds or less?’ Driving down the highway, a glance at your dashboard tells you everything you need to know,” Ohlund says. “If you’re designing displays that take minutes for a human to analyze and digest, then functionally it’s not a dashboard.”
Armada’s Data Projects
Armada plans to once again double its available storage space in order to acquire point-of-sale data and explore its connection to the larger supply chain.
Armada expects to see all trading partners relaying data on a near-hourly basis within two years, and in real time within ten years.
Currently, dashboards must be consulted in order to identify hiccups within the supply chain. Armada would like to program the system with data triggers that will recognize potential faults and alert trading partners via e-mail, instant message, or a phone call.
Currently, many system processes are still primarily dependent on human operators, but that is expected to change. “We’re establishing business rules and manual processes with the intent of automating and codifying those business rules so that we can automate the responses to data triggers,” Ohlund says.
Think about those premium freight charges. Under the current model, the system generates a report indicating the shipping space required, but is still dependent on an operator to act on the report. The next step will be creating proactive alerts to notify an operator when reserved shipping space is below the required threshold. The culmination will be an entirely automated process: a set of data triggers automatically tenders a request to the carrier. “Each time you can define the limits of an operation and codify what you do to remediate occurrences outside those limits, you automate more and more of the supply chain,” Ohlund says. “The human operators involved become the detailed exception handlers, and the management and operation of rote supply chain activities become increasingly efficient through automation.”