A private or dedicated fleet can solve a lot of your distribution and fulfillment challenges, especially if your business runs on regular or time-sensitive deliveries (e.g., retail, DSD, QSR, etc.). As the saying goes, consistency is key. Unfortunately, this operational elixir has some unintended side effects when it comes to strategic network modeling. Let me explain.
If you’ve ever built a supply chain network model, then you know that developing transportation cost inputs for a private or dedicated fleet is much harder than gathering cost inputs for common carriers operating point to point. Unlike point-to-point modes, where historical cost is readily available at the shipment or lane level, transportation costs for a private and dedicated fleet are generally not available at this level of detail. Private fleet costs are typically recorded in monthly increments in a general ledger format at a domicile level. Similarly, dedicated fleet costs are typically invoiced weekly at a domicile level. In both cases, costs are not available at the shipment level or lane level (i.e., the required level of detail to build a network model which builds confidence in the recommended design).
To further complicate matters, many private and dedicated fleets are used for multi-stop deliveries. Therefore, the cost of an individual delivery isn’t just dependent on the distance from the origin to the delivery stop and the size of the delivery; it’s dependent on a variety of factors including the configuration of the delivery route, the time required to execute the route, and the utilization of the delivery equipment.
Unfortunately, because network modeling software only thinks in terms of point-to-point lanes, network design practitioners are forced to somehow create lane-level point-to-point cost for fleets making multi-stop deliveries. Most network design modelers employ one of two techniques when developing transportation cost inputs for private or dedicated fleets:
- LTL Rating: One of the most common techniques is to treat fleet deliveries as LTL shipments. Essentially each individual delivery or average delivery is rated using a scaled LTL tariff, setting the scaling factor such that the total rated fleet expense is equal to the total actual fleet cost. Once this scaling factor is determined, the LTL tariff can be used to rate any fleet activity in the network model.
- Cost per Unit-Mile: The second most often used technique involves developing a cost per unit-mile. This can be done by summing the product of all units delivered and the origin-destination distances over which those units were delivered and then dividing the actual fleet cost by the sum of those unit miles. Once this cost per unit mile is determined, it can be used to rate any fleet activity in the network model.
Although these approaches resolve the challenges associated with transforming cost resulting from executing multi-stop deliveries into lane level point-to-point costs, and both approaches can be used in conjunction with cost information available in either general ledger or dedicated invoice form, we’ve actually found that the resulting transportation cost inputs from those two approaches are grossly inaccurate. And it’s the reason we developed a new approach to fleet costing which considers distance, time, utilization, and multiple stops.
As shown in the above charts, comparing both the LTL and cost per unit-mile approaches to our proprietary fleet costing approach for three shippers operating fleets with three different average stops per dispatch, the common approaches overestimate cost for routes with shorter stem distances and underestimate cost for routes with longer stem distances. These errors appear to be even more pronounced on shorter stem distances for fleets with fewer stops per load and more pronounced on longer stem distances for fleets with more stops per load.
The common approaches [to fleet costing] overestimate cost for routes with shorter stem distances and underestimate cost for routes with longer stem distances.
Put another way, the tendency of both common approaches is to underestimate the cost of private or dedicated fleets the greater the distance from the origin and/or the greater the number of stops per load. Using transportation cost inputs like this could wrongly suggest that fewer fleet domiciles would be less costly to operate than it would actually cost if implemented. And even worse, this error would only be found once you’ve implemented the design and incurred a higher cost than expected, eroding confidence and executive support.
If your organization is struggling to develop a fleet costing approach that you can trust for investment-grade network modeling, our supply chain network design experts are eager to help you understand the alternatives and determine the best strategy for your company.
Jeff Zoroya is a supply chain strategy expert with 20+ years of experience from every angle – consulting, software, and industry. At Chainalytics, Jeff leads data-driven engagements related to supply chain network design and strategic planning across all major industries, including Retail/Wholesale, Food & Beverage, Consumer Goods, Hi-Tech, and Industrial.
