From now until Christmas, UPS is in peak season. This year the international package-delivery company expects to deliver more than 700 million packages by the end of December, a 14 percent jump in deliveries compared to the same time last year. To meet U.S. demand, the Atlanta-based company expects to hire 95,000 part-time and seasonal workers. To save mileage and money along the way, UPS is relying more on big data and a tool it calls Orion.
“At UPS, we always say we’re an engineering company with technology — we just happen to have planes and trucks as well,” said Mark Wallace, senior vice president of global engineering and sustainability.
How to best deliver the 18 million packages UPS ships daily is the biggest question the company faces. Drivers of the company’s signature package trucks make about 120 stops per day and face seemingly endless alternatives when it comes to the order in which they make deliveries. Now, using a combination of big-data analytics and algorithms, UPS is able to track packages closely enough to let customers know when to expect a package and guide drivers out on deliveries into choosing the best route.
It’s a challenge the company has faced after hitting snags in 2013 and 2014 as e-commerce holiday sales have skyrocketed. This year volume should peak. The National Retail Federation expects 2016 online holiday sales to increase between 6 percent and 8 percent.
The Orion computer platform, a 1,000-page-long algorithm more than a decade in the making, is the biggest piece of the puzzle. Developed by a team of 50 UPS engineers in Maryland, piloted through the end of last decade and launched in 2013, Orion stands for On-Road Integrated Optimization and Navigation. As a UPS delivery driver makes their rounds, Orion works in the background. considering up to 200,000 route possibilities before picking the most optimal path from one delivery stop to the next and reordering a driver’s delivery schedule if necessary.
“What Orion is doing is reducing miles for us,” said Glenn Zaccara, a UPS spokesman. “The focus isn’t so much on reducing the time to deliver but rather on reducing the number of times a package is touched by a human in our system, or the total amount of time that a package spends in our facilities.”
In this way, the information that flows with each package is as important as the package itself and helps not only drivers but customers. My Choice, launched by UPS in 2011, sends customers alerts on when to expect a package delivery. But if one of the 30 million UPS My Choice members uses the platform to change the time or location of their delivery, Orion will recalculate the driver’s route to accommodate the change, and do so in less than 10 seconds.
The growing reliance on data analysis by UPS follows a larger trend in the transportation industry, where telematics and analytics are of growing importance in fleet management.
“Now companies can collect unstructured data, which was not economical to store, as well as historical data,” said Gartner analyst Svetlana Sicular. “Now that you can store more data, you can see longer-term patterns. … The goal of this is to make decisions more consistent.”
During peak season, consistency is key for UPS. In 2013, thanks to snow and a deluge of last-minute, online orders, an estimated 2 million packages went undelivered by Christmas by both UPS and FedEx, but a majority of those packages were in UPS’s network. The problem was solved in 2014, but at a cost to UPS, which had to lower its profit forecast for the 2014 fiscal year.
Even more pressing for UPS is the changing nature of the delivery business. The e-commerce market — which surpassed $340 billion in 2015, according to the Commerce Department, and are on pace to equal or surpass that level this year — ensures that more packages are delivered straight to customers’ doorstep. This is a challenging evolution over the last 20 years for UPS and other delivery companies used to dropping off groups of packages at retailers.
Orion is meant to cut costs and make more efficient use of time as drivers are being called upon to make more stops. So far UPS has spent $250 million this decade on Orion integration across the company’s 55,000 U.S. delivery routes. When full Orion integration happens next year, UPS expects to save between $300 and $400 million per year thanks to the algorithm, according to the Wall Street Journal.
“At UPS we always say we’re an engineering company with technology — we just happen to have planes and trucks as well.”
In some ways, UPS saw these changes early on. Starting in the 1990s UPS began streamlining operations, first by installing GPS-tracking equipment on delivery trucks and replacing handwritten delivery slips with “smart labels,” bar-coded slips that are scanned as much as five times as packages move from processing facilities to delivery trucks to the hands of the customer. Trillions of pieces of data are collected along the way, which inform the Orion algorithm.
It also embraced telematics early. UPS package trucks are equipped with embedded sensors to check the health of the truck’s battery and count how often a driver backs up during a workday. Backing up is a key metric for determining a driver’s likelihood of getting into an auto accident; for drivers who back up too much, supervisors will work with them on driving strategies to reduce the number of times they go in reverse.
How revolutionary Orion will be for UPS is what the company is assessing now. According to UPS, the algorithm is saving the company $300 million to $400 million annually by helping drivers shave between six and eight miles off their routes per day.
UPS is still refining Orion to strike a balance between helping a driver find the best route and maintaining delivery consistency. A business used to receiving morning deliveries probably won’t want late-afternoon deliveries because an algorithm told a driver to consider an alternate route.
But as the company’s data capabilities get more sophisticated, UPS will try to make the Orion algorithm more predictive. While Orion finds the best routes for drivers, it doesn’t yet take into effect things like traffic conditions that can change at a moment’s notice, nor does it offer turn-by-turn directions like a traditional GPS system.
“The next versions of Orion, we’ll go into a dynamic version that really will start to take into effect things that do happen to delay drivers potentially on the street,” Wallace said. “That is going to be taking place over the coming years across the globe.”
UPS will deliver 700 million packages during this year’s peak holiday season. Worldwide, the company delivers 18 million packages daily.
Every delivery driver makes on average about 120 stops.
The Orion algorithm optimizes driving routes by choosing the best route for each delivery driver to take, shaving off roughly six to eight miles of drive time per driver.
So far, Orion has saved UPS 180 million miles and 10 million gallons of fuel.
By 2017, UPS expects drivers along all 55,000 of its North American routes will be using Orion.