Editor's note: This article is the first in a series that looks into the ways supply chains, warehouses and manufacturing facilities are investing in technology. Here's the next story.
Predictive analytics isn't a new technology. In fact, it's older than the United States. In 1689, Lloyd's of London starting using predictive analytics to underwrite sea voyages.
Since then, predictive analytics has launched ahead as access to information became better and faster, with computers giving a major assist.
In a modern sense, "they're similar to RFID tags. They've been around for a while," said Rudolf Leuschner, associate professor in supply chain management at Rutgers Business School. "When we look at the actual algorithms that underpin predictive analytics, there isn't really that much new stuff."
What has changed is the sheer amount of data, and what's done with it.
"We've gotten to the point where people aren't screaming that we don't have enough data. They're screaming about how we have too much data," Leuschner said. Software and technology companies are "doing a better job at taking in that fire hose of information and making something out of it," he said.
A ubiquitous expectation of real-time ETAs
In "Innovation Driven Resilience," the 2021 MHI Annual Industry Report, MHI and Deloitte surveyed more than 1,000 supply chain professionals worldwide about innovation investments in the supply chain. They found that 31% of respondents say predictive analytics are already in use, and 48% say it will be in use in the next five years.
The success of predictive analytics has already changed expectations.
"We've gone from almost having zero expectations around things like ETA and what's the likelihood that a shipment's going to show up in the timeframe it was originally predicted to narrowing that down," said Ken Wood, executive vice president of product management at Descartes.
That expectation is fairly ubiquitous now — for shippers, forwarders and receivers, whether that receiver is a business or a consumer.
"Everybody's expecting to have increasingly narrow guidance when they can expect the goods to arrive," Wood said.
Planning, forecasting among top uses for predictive analytics
The technology has also created close to real-time visibility in the most advanced supply chains.
"Now we're on our way to, you can see where the shipment is relative to your location for the last 30 minutes," Wood said.
The widespread adoption of cloud has helped predictive analytics along, he added, because they provide easier access to machine learning algorithms and big data processing.
"People aren't screaming that we don't have enough data. They're screaming about how we have too much data."
Rudolf Leuschner
Associate professor in supply chain management at Rutgers Business School
The better technologies get at sorting through data, the better predictive analytics will become, especially when it comes to dealing with uncertainty, said Wood.
"Predictive analytics work best when they're making predictions in stable circumstances," he said. "But when the extraordinary happens, and some of that stuff is driven by external factors, whether it's weather or labor actions or geopolitical struggles, [the technology] isn't as good at making predictions in those kinds of scenarios."
Cost barriers of predictive analytics
Predictive analytics is still an expensive technology to set up. Even if it can save costs once up and running, not every enterprise can make that kind of capital expense.
According to the MHI report, more than half of enterprises are spending $5 to $10 million on predictive analytics. Three percent are spending more than $100 million.
Companies spend big on predictive analytics
For some warehouse operators, that kind of expense doesn't make sense, said Leuschner. Adding predictive analytics to a warehouse can make workers slightly more efficient and slightly less error prone, but unless the company is operating their warehouses at scale, the savings won't be realized at these price points.
That could change if the systems can be installed as an app on a smartphone without much maintenance involved, instead of requiring investment in hardware, software and consultants to make it all work. according to Leuschner.
"If a company can load an app on my phone and build it on a per use basis, that's a lot easier to implement," Leuschner said.
This story was first published in our weekly newsletter, Supply Chain Dive: Operations. Sign up here.
