Background noise is the drip-drip water torture of modern supply chains.
And the impact is felt no more keenly than in transportation procurement. This is where the overused concept of big data becomes a tangible, visceral problem for transportation managers in the digital age.
Buyers of transportation services are literally inundated with information these days. That’s a good thing, since more information leads to better decisions. Unless the information is so voluminous, uncategorized, and disorganized as to make the decision harder.
Think of it like this—imagine you are an extra-terrestrial landing on earth and one of the first things you see is a supermarket full of shelves of food and household items. If you don’t even know what food is, or what a home is, how would you begin to make sense of the minute differences between all the products offered? What makes 1 percent milk different from 2 percent milk?
And so this is a long way of explaining that transportation buyers have a big challenge in terms of managing all the data points coming their way. But you probably already knew that. If you’re buying ocean, or trucking, or intermodal, or air freight capacity in any significant amount, you know it’s a juggling act, balancing the traditional means of negotiating contracts with more modern forms of automated bidding.
So here’s where I’m going with this. There’s a prevailing sense that technology is the fine point on a freshly sharpened pencil, or rather the tool that makes the pencil point fine. But is that really the case?
“There’s definitely the belief that if you use bid optimization tools, you hit the magic button and you get the answer,” said Kevin Zweier, vice president of transportation practice at the supply chain consulting firm Chainalytics. “There’s certainly more to it than that. And even the next step, where they know it’s about pressing a button, they still think, ‘I want the tool to tell me the things to look at.’ Well, they don’t really do that either.”
Instead, optimization tools take pre-set constraints and work to uncover every shred of inefficiency and extra cost.
“Optimization is greedy,” Zweier said. “It finds all the places where it can shave off those costs like crazy. But the problem is, a lot of times, that doesn’t jive with what the shipper actually wants to do.”
Indeed, it’s better to think of transportation systems as tools that empower decisions, not decision-making tools.
“This is not an indictment of the tools themselves,” Zweier said. “These tools are fantastic at what they do. But a client needs to know their network and where their constraints are. In the truckload world that means they need to know, for instance, how much intermodal a certain provider can actually do. The tools can’t tell those things. They can test those things, and that’s where the power is for sure.”
Shippers that already successfully use bid procurement tools, or seek help from folks like Chainalytics to do so, understand that the system is not the decision-maker.
“Even a customer who knows the types of constraints they want to evaluate, what we find in every project, regardless of mode or tool, is that scenarios will only get you 80 to 90 percent of the way there,” he said. “It’s unrealistic to think you’ll get the whole way.”
Much of this is down to practicality. It would literally take too much time, Zweier said, if you went lane by lane expecting your procurement tool to make decisions.
“To code that to that degree is like chasing your tail,” he said.
For instance, a bid optimization tool may say that carrier “X” should win a contract for a specific lane. But the traffic manager in the region for that lane might know that this particular carrier has difficulty providing capacity on that lane.
The system doesn’t know that carrier “X” struggles in that lane because that minute piece of information was never inputted as a constraint in the optimization engine. And on the other hand, that regional traffic manager wouldn’t even think to input the reality about carrier “X” into a system that accounts for that information because he’s not been trained to do so. Bid optimization is generally a centralized exercise.
So the process, in reality, works like this: carrier “X” wins the bid for that lane, but the regional traffic manager shoots down the bid and says go with the second bidder because they’re more reliable in that lane.
The crucial nature of timing in procurement events would trump any desire to make it an entirely automated process.
“Human judgment comes in, because the tools have to obey the constraints presented to them, and humans want to bring that judgment in,” Zweier said. “Every client we work with does some level of adjustment after running scenarios. Almost every customer I work with wants to touch and review those adjustments at the end. And they’re not disappointed about having to do that. I’m not sure how many would trust $100 to $200 million in transportation spend purely to an optimization engine.”
This concept doesn’t apply purely to transportation procurement. Nearly every system that automates functions across the transportation and trade environments are designed to allow decision-makers to make better decisions, not to replace them.
“Same principle applies,” Zweier said. “You want the technology to take care of the easy stuff, so you can focus on exceptions.”
Filtering out the noise—that’s what technology brings to transportation decision-makers. And the more modes you manage, the more geographies you touch, the more data to synthesize, the more important that filter becomes. Just remember, it’s humans that still drive the bus.
This column was published in the March 2015 issue of American Shipper.