When the business gets too complex for spreadsheets, capacity decisions need a different foundation: one model of the operation that gives provably optimal answers and shows why. Here is how FedEx Office built that foundation.
Defining the Objective
FedEx Office is comprised of nearly 2,200 retail stores, 15,000 employees, 18 commercial print facilities, and a nationwide transportation fleet. Tom DeGreve, Managing Director of Production Strategy & Engineering for FedEx Office, with advice from Pierre Mawet, Managing Director of Supply Chain & Operations at Accenture NA, set out to better understand three things: the complexities of print for the company's customers, the costs of transportation and production, and how best to meet demand at any given time. Better insight into these areas would drive operational improvement and resource adaptability, leading to stronger customer relationships and a better business overall.
Finding the Right Approach
Initial optimization efforts at FedEx were narrowly focused. The starting point was a relatively small, internally driven, loosely structured approach that took a great deal of effort and explored several simulation-based methods. As the team added more variables and complexity to their models, assessing them in Excel and the other tools in place became increasingly difficult and impractical. That difficulty prompted the search for a better approach.
With the Enterprise Optimizer® (EO) Decision Engine, FedEx Office could build a constraint-based model of its full value chain. The model would include all key functional components (commercial, operational, and financial) so the team could assess a virtually limitless range of variables and options. After gaining internal buy-in, the company moved forward with the Decision Engine to create a capability it could build on and draw more value from over time.
Biggest Challenge
Some organizations don't have enough reliable data readily available to realize the full power of the Decision Engine. Their data may be aggregate-only or unvalidated, so it won't accurately capture the interplay between capacity, assets, labor planning, and variable costs. FedEx Office had the opposite problem. They initially loaded so much data that it crashed their models. This prompted important conversations about, as DeGreve explained, "how we think about the business." In short, the team needed to find the right level of detail without settling for a rudimentary aggregate.
With the Decision Engine, FedEx Office now has a model that continuously runs a 10-day view of baseline down to the tens of thousands of interactions that take place every week. This gives a highly detailed view of behavior across the whole system. The model also accounts for seasonality, time elements, and other constraints to show the best way to service demand at any given time. It is a true model of operations, not just an aggregate for a certain period.
From Insight to Action
Before the Decision Engine was in place, decisions on service components such as order routing were guided by mostly absolute policies. Now FedEx Office can identify conditions where deviating from those policies is warranted to achieve higher profitability and a better customer experience. For example, current conditions may show that delivery by 3 pm vs. 5 pm makes a significant difference in costs and resources. If 5 pm delivery is acceptable to the customer, they can get a lower cost for the same quality service.
FedEx Office first pilots policy adjustments like these in a few markets, then expands them as they prove their real-world value. In the short time the Decision Engine has been active, they've realized a solid ROI that will only grow. In a webcast, DeGreve cited one example, saying that "delivery from plants and back to stores is a significant annual spend – one we now believe we can cut by up to 45%." That's a big number, and it covers only the transportation portion of the service equation.
A Foundation for Continual Optimization
FedEx Office is one of many organizations using River Logic's decision engine technology to drive breakthrough improvements in capacity planning, supply chain network design, business continuity planning, and more. Because the model is built once, each new planning question builds on it instead of starting from scratch.
Moving forward, these organizations are positioned to unlock the enormous pent-up knowledge held by their most important asset: their people. The Decision Engine delivers answers quickly and clearly, with the reasoning behind each one. That encourages the creativity and communication that bring crucial questions to the surface, questions that lead to healthier businesses and stronger partner and customer relationships.






















