Increasingly, data can tell sophisticated stories that help businesses improve their operations. But all data isn't created equal. How can LTL leaders know whether they're asking the right questions? And how can they make sure that the questions they're asking aren't limited by a myopic perspective?
These questions anchored a recent session of SMC3's LTL Online Education Hybrid Series on business analytics. The session was hosted by SMC3 and Dr. Karl B. Manrodt, a Professor of Logistics and Transportation from Georgia College and State University. Featured speakers included Scott Friesen, Executive Vice President for Strategic Analytics with Echo Global Logistics and Mike Marek, Director of Strategic Pricing at Worldwide Express.
The central question: "Compared to what?"
To be effective, data needs context. Through context, organizations can build a better sense of their data's significance and identify the kinds of stories it can tell. Friesen noted that if you could ask only one question within business analytics, it should be, 'Compared to what?'
For example, you might say a certain condition has increased by 50% in your business. But if the initial base rate was, for example, 1 out of 100,000, that would only increase it to 1.5 out of 100,000. Raw figures don't necessarily tell important stories -- they need a baseline of comparison to pull out their true importance.
Having these comparative variables is central to making data useful. And, in many cases, organizations just won't have the information they need to answer the questions they're curious about.
"Sometimes you have it and it's hard to work with," Friesen explained. "Sometimes it doesn't exist."
Looking outside the industry for answers
In leveraging data to gain insight on organizations, leaders often fall prey to preconceived answers. These preconceptions can create tunnel vision, obscuring the bigger picture, and, more importantly, overlooking opportunities to improve flawed processes.
To counteract this tendency, LTL leaders should be in constant dialogue with people outside their industry. These cross-sector conversations can help facilitate crucial knowledge transfers. For instance, truckload is currently dealing with fraud problems that have long-plagued banks and retailers. Seeing how these sectors have responded can help trucking figure out if its own best practices are sound.
Not only are these conversations helpful for finding new answers to vexing questions, they also help LTL leaders clarify their existing problems and solutions by forcing them to explain them to people outside the industry.
"As someone who came from retail, the idea that an LTL company would not know its win rate made my head absolutely spin," said Friesen. "Every store manager knew their conversion rate every day. The idea that I stepped into an industry where that was just not broadly known was amazing to me. Then we went about and fixed it."
A new ability to predict
The last five years have brought a new awareness in LTL that data science can actually predict outcomes and answer meaningful questions within the industry. Previously, leaders wouldn't ask questions of data scientists simply because they didn't know that they could.
Predicting whether an LTL load would get delivered on time was formerly considered impossible, for example. But increasingly, companies are coming to data scientists to model these scenarios. Thanks to the proliferation of more data, these predictions are becoming more accurate.
AI on the horizon
AI has also taken entire classes of data and transformed them from being prohibitively expensive to becoming both feasible and profitable. AI models still need refinement, and the industry is captured simultaneously by AI hype and disillusionment. Yet it's clear AI will eventually be a significant piece of business analytics for LTL.
Marek noted that initial use cases would probably center on efficiency improvements for low-hanging fruit like back office administrative functions, operations, and sales.
Three elements of analytics
Friesen listed three core pieces to business analytics: observing, relating, and predicting. Each of these help improve enterprise decision making. New AI models at a minimum will help enhance the relating function, painting a more complicated relationship between variables as they start to unlock the complexity of unstructured data.
Stay curious
As new technologies advance at a rapid pace, and LTL continues to improve its reliance on data science to more accurately define where it's going, both panelists encouraged LTL professionals to stay open-minded about how the industry can advance. This openness doesn't just mean building strong relationships inside and outside the industry, but it also means developing a voracious habit of learning to pull in new solutions.
"There are lots of opportunities if we just break down silos and collaborate," said Marek.