Alan Robinson, President, Direct Communications Group, is one of our guest bloggers in the “Five Elements of a Postal Solution” blog series. This is the full version of Alan Robinson’s blog.
In trying to define an optimized Postal Service infrastructure, I was stuck at trying to define what “an optimized Postal Service infrastructure” meant. The problem that I had was not with the mathematics of optimization which is focused on either cost minimization or profit maximization but with four sets of constraints that affect the formulation of an optimization model. These constraints are:
• Customer service needs which determine product characteristics;
• Labor management constraints that impact operating costs;
• Capital spending constraints that limit options for infrastructure optimization; and
• Regulatory constraints which modifies a market-driven optimal network to reflect political considerations.
While in the future, the Postal Service’s customers are going to need a postal infrastructure that offers both physical and digital delivery, this post will focus on the physical delivery infrastructure. Information infrastructure will be discussed only to the extent that an information infrastructure is needed to manage a physical infrastructure or provide product features that are a component of a physical delivery product. Information infrastructure that would be needed to either provide coordination between digital and physical delivery of different documents or transform a digital document into a physical one or visa-versa that an optimal Postal Service should offer will not be discussed.
Customer service needs determine product characteristics
The service needs of the Postal Service’s customers have six dimensions:
1. Access to services in a convenient manner;
2. Delivery to a location that satisfies the needs of recipients documents and parcel;
3. The speed that a document or parcel must be delivered from point A to B;
4. The reliability of a promised delivery commitment;
5. Information on delivery status and reliability provided on a real time or near real time basis; and
6. A price that makes the service financially viable.
The discussion below reflects current customer needs. However, customer needs constantly change. Communications access, processing, transportation, and delivery infrastructure will need to be re-optimized constantly.
For the Postal Service’s communications customers (i.e. bulk mailers), the existing network will likely need to adjust over time to improve the ability of Postal Service customers to coordinate delivery with communications sent digitally. This will likely require shorter cycle between when a physical document begins printing and when it is delivered.
For small parcel shippers (i.e. parcels under 10 pounds), the real challenge, is having a parcel delivery network that can assist in reducing the time from order-to-delivery, handle volumes growing by 20% or more annually, and provide the service at prices that allow the e-retailer to offer free shipping to their customers. This will likely require the Postal Service to constantly adjust its parcel processing and transportation infrastructure to reflect growing volumes while seeking new delivery methods that offer both secure and inexpensive delivery.
The issue of access is usually focused on retail access but includes access to a delivery network for bulk mailers and digitally created documents.
An optimal retail network has to have access points determined primarily based on customer demand tempered by USO requirements to serve rural and lower-income communities. The optimal network can have any combination of corporate and contract offices. Based on international experience, it appears that an optimal network requires that individual locations have the commercial freedom to offer any product or service that can generate revenue to sustain a retail enterprise in addition to postal services. This generally means that most retail locations are contract locations but corporate locations are an option if they are given the needed commercial freedom.
The access infrastructure will likely need to include self-service options such as GoPost. While useful for general parcel shipping, the growth in single-piece parcels will likely come from recipient-paid parcel returns and a low cost acceptance location is critical both to handle the expected growth in volume and e-retailers need for a very low cost return option.
For bulk mailers, access is determined by the processing and delivery network. Within an optimal network, bulk mail access will occur only at places which fit the Postal Service’s operating plan. This will likely include only processing plants and delivery units. Bulk mail access at other locations will need to have prices that include the additional transportation and handling costs to move the mail to or parcel to the nearest processing plant. Optimal access for bulk mailers will remain constant only as long as the processing and delivery network does. Optimally, acceptance locations will change whenever processing and/or delivery activities change their locations.
The optimized physical delivery infrastructure will need to include two delivery options. These options are:
• Door, curbside or cluster box delivery for most items and only door delivery for signature items; and
• A retail outlet or parcel locker that provides parcel pick-up for customers that are not home for delivery.
The Postal Service has proposed shifting some door delivery to curbsides or cluster boxes and some curbside delivery to cluster boxes. It is clear that cost and service are important for both document and parcel delivery. Alternative delivery firms have prospered in door delivery when either the Postal Service’s service quality is too low or existing prices are too high. While moving delivery away from the door is cheaper, it is unclear to me whether shifting delivery location to reduce costs is the optimal solution to serving sender needs and ensuring financial self-sufficiency. However, other constraints affecting the ability of the Postal Service to control costs or set prices may force a sub-ideal solution regarding delivery location.
Regardless of the delivery location, the delivery infrastructure must have an information infrastructure that allows for route adjustments on a frequent basis. Ideally, delivery routes need to adjust to reflect seasonal changes in mail volume. As parcel shipments increase, parcel routes may become more common and these routes need to be able to be adjusted on an even more frequent basis.
The key to an optimal parcel delivery network in the future is finding ways to cut costs so that e-retailers can offer free delivery on more sales as well as finding a way to provide delivery to recipients that are not at home during normal delivery hours. Therefore, an optimized Postal Service’s delivery infrastructure will also need to have delivery locations that permit recipients to pick-up their parcels at the recipient’s convenience. While Post Offices are an option, limited hours suggest that a parcel locker system like GoPost could more easily meet both recipient need for access to their parcel shipments on nearly a 24-7 basis and e-retailers need for a low-cost delivery option that is secure. Given the expected growth in small parcel delivery volumes, and acceptance of similar lockers in other countries, an optimal GoPost network could include ten to thirty thousand locations in order to meet both shipping volume and recipient demand.
Delivery Speed and Reliability
Both the current network, and the one that the Postal Service plans to implement in May make assumptions about the delivery speed required by customers. The current network meets the expectations of most single-piece mailers but has difficulty consistently meeting the expectations of many bulk mailers and its delivery standards for Priority mail are not competitive for most short distance shipments.
The Postal Service’s proposed network meets the current delivery speed expectations of about 75% of its customers and arguably can meet those expectations with more reliability than the existing network. It cannot meet the current service levels of single piece mailers, some periodical mailers, and would likely have even more challenges meeting expectations of Priority Mail shippers. While problems with meeting the expectation of periodical and Priority Mail customers are known, little is known as to whether the new slower service standards meet the expectations of single-piece mailers or not.
Ideally, the Postal Service would have the option to maintain the current network and service standards by charging single-piece and Periodical mailers the additional cost associated with the infrastructure needed to provide current service levels. However, current law and regulatory precedent would not permit the large increases in single-piece First Class mail and Periodical rates that would be required even if customers would be willing to pay the higher rates for maintaining current service levels. While the willingness of First Class mailers to trade higher rates for maintaining current service levels is unknown, periodical mailers would likely find the rate increases required could force magazine subscription rates above levels that make mail-delivered subscriptions a profitable business.
Given that pricing adjustments are off the table, setting an optimal network requires making some choices among customers. The Postal Service has little chosen but to focus on the current expectations of bulk mail and Priority Mail customers and fit service levels provided to other customers around what that network can provide. This makes sense given the expectations of volume growth (or losses) over the next decade for the full range of current Postal Service products.
Regardless of decisions to make a processing and transportation network as close to optimal today given other constraints, that network is unlikely to remain optimal for long. Both mail volumes and current expectations of delivery speed for mail and parcel shippers today are unlikely to reflect the volumes or expectations of Postal Service customers 10, 20 or 30 years from now. This means that the optimal infrastructure needed to meet customer expectations while minimizing costs will change over time.
Right now it would appear that an optimal infrastructure may need to adjust to mailers’ interest in reducing the cycle time from when a document is printed until it is delivered in order to coordinate with digital communications and the needs of parcel shippers interest in reducing the time from when an order is accepted until delivery to the preferred delivery location is made. Reducing delivery time for mail and parcel will require a larger infrastructure but also changes in labor contracts and availability of capital to make sure that the reduction in delivery time can be accomplished without unacceptable increases in rates.
Delivery and Management Information
The Postal Service’s infrastructure needs to provide real-time or nearly real time information on delivery status and reliability. While the Postal Service has developed systems that can provide this information, capital constraints have slowed deployment. As mailers expand their coordination of physical and digital communications, this information will be increasingly important and delays in implementation create the risk that the Postal Service’s delivery of physical documents will become even less competitive.
For Parcels, the problem is even more important. The Postal Service has become a critical link in e-retailing. It now delivers nearly 30% of FedEx Grounds parcels and UPS has just begun promoting SurePost service. Both carriers expect that the Postal Service will handle most of their parcels under 10 pounds. In order to meet this demand, the Postal Service’s parcel tracking system needs to become fully integrated with the systems of FedEx, UPS, DHL-Global Mail, Newgistics and any other firm that wants to use its delivery network. This includes integration of whatever information system is used for GoPost delivery lockers as well.
From a management standpoint, the information needs of the customers fit right into the information needs of management. Management information infrastructure must have the capability to shift operating flows on a real time or nearly real time basis to prevent network bottlenecks and to monitor and adjust delivery networks to minimize cost and maximize service on nearly a real time basis.
Mailers budgets currently reflect existing postage prices. This means that the Postal Service’s ability to provide a certain level of access, a particular delivery method, and a service speed is limited by the price that its customers pay and the need to ensure that prices provide a sufficient return to cover operating costs, legitimate long-term liability obligations and sufficient cash for future emergencies and capital investments. This forces the Postal Service to make choices as to how it provides access and delivery, and the speed of service that it offers the various market segments that it serves. The optimal access, delivery, processing, transportation and information infrastructure therefore is constrained by the prices that the Postal Service can charge.
In the future, changes in prices for all postal services will affect demand for the product and the infrastructure needed to provide each service. The impact on volume of prices changes will vary based on the sensitivity of a product’s customers to price. For products like First Class single-piece mail, the impact of price changes on volume and therefore the infrastructure needed to handle the volume is small. For other products, the impact of price on volume and the Postal Service’s infrastructure needs is likely larger. For all products the actual relationship between price and volume is less clear than it has ever been as the growth of competitive web-based and mobile-based communications alternatives have created significant non-price related reasons for shifting from postal to digital delivery.
Labor-Management Constraints Affect Infrastructure Decisions and Service Quality
As labor costs will always represent the majority of operating costs, any deviation from competitive levels of employment costs raise the cost inputs of optimization models and force an optimization modeler to generate a sub-optimal infrastructure to serve customer needs at prices the Postal Service can charge. Currently the Postal Service’s greatest constraints affecting managing the labor component of its infrastructure include:
• The legal obligation to continue to fund pension accounts (i.e. FERS and CSRS) that are already overfunded;
• The legal obligation to fund a retiree health benefit account on a schedule that is not based on a current estimate of the underfunding and on a pace that forces significant financial losses at current rates;
• Inability to control health and other benefit expenses by offering a benefit package designed for postal employees and not all Federal employees;
• Restrictions in offering retirement and other incentives to adjust the size and skill mix of the workforce; and
• Labor contracts that restrict flexibility in regard to part-time and full-time employees, and adjust schedules based on changes in seasonal demand.
Networks designed under current contracts must take into account the increased costs and reduced flexibility that these constraints create. All of the constraints that raise the Postal Service’s average hourly employment costs raises the total cost of the network forcing the modeling exercise to develop the network that provides a service quality less than would could otherwise be provided given the price and therefore the overall cost constraints. Limitations on full and part-time employees, as well as seasonal adjustments to schedules force the modeler to identify operating plans that maximize the use of full time employees in order to minimize cost, even if those operating plans result in a lower service level than an operating plan that can incorporate a higher proportion of part-time employees could offer.
Removing or relaxing constraints imposed by law or contained in existing labor contracts would affect decisions about the size of the network infrastructure that the Postal Service deploys. Lowering labor costs would likely increase the size of the optimal Postal Service infrastructure as the lower labor cost would allow the development of operating plans (which affect the infrastructure within individual plants and used in retail and delivery operations) and the number of Postal Service facilities as they allow the Postal Service to provide a higher level of service within a price constraint.
Removing or relaxing In the future, customers are likely to require an infrastructure that can provide a physical delivery service that can provide service at current or near current prices, as adjusted for inflation, with shorter cycle times than now are possible at current prices in order to work with the short cycle times available in both digital and broadcast media. This need will likely require a larger infrastructure than the Postal Service envisions in its network optimization initiative. However, these future customer needs cannot be met as long as constraints imposed by law or contained in existing labor contracts remain.
Capital Constraints Result in a Sub-optimal Infrastructure
The issue of capital constraints on the Postal Service infrastructure can be illustrated by the network optimization that the Postal Service plans to implement in May, delays in implementing its information architecture, and delays in replacing its delivery fleet. In all cases, capital constraints have resulted in a sub-optimal solution.
The Postal Service’s network optimization used only existing facilities and existing equipment. A green field optimization would have continued to use many existing plants but in many parts of the United States a new or expanded plant would have provided the proposed service commitments at a lower cost or even better service commitments at the same or lower operating costs than what were possible using only existing facilities. The issue of capital constraints also reflects the need to keep certain annexes open as a more optimal solution would merge operations now conducted in the primary facility and the annex into one facility but the primary facility is too small or has other problems that force keeping the annex open.
The problem with capital constraints has also affected the Postal Service’s ability to have optimal information architecture. Systems that track mail and parcel location, manage delivery routes, manage vehicle routes, manage plants and ensure accurate information on in plant volumes and costs have seen budgets cut and project schedules extended by months and in some cases years. As such, the Postal Service has fallen behind current needs of its customers and more than likely is not near having an optimal information infrastructure.
The problem of capital constraints has affected the Postal Service’s replacement of its delivery vehicle fleet. Fleet replacement is a multi-year, multi-billion project for a fleet as large as the Postal Service. Not having the capital to replace the fleet increases operating expenses for repairs and maintenance and increases risk of service failure due to vehicle breakdowns. Not having the capital to replace the fleet may even influence decisions over 5-day delivery as eliminating one delivery day reduces vehicle use and may extend the life of the existing fleet.
The full impact of capital constraints on the Postal Service infrastructure today is not known. Clearly it affects both operating costs and service quality and makes it difficult for the Postal Service to provide the information interfaces that customers demand.
As customer needs change, capital constraints will continually ensure that the Postal Service will have a sub-optimal infrastructure to meet those changing needs. This will most likely result in Postal Service operating costs (and therefore modeling costs for designing the optimal infrastructure) being higher than they would otherwise be, and the Postal Service will have a less extensive infrastructure than makes business sense. In particular, I am concerned that capital constraints will make it difficult for the Postal Service to adjust the network in order to improve service or reduce costs and have the information, facility, equipment and vehicle infrastructure to meet future customer needs for shorter cycle times at or near current prices (as adjusted for inflation) for postal services.
Regulatory Constraints Affect How Optimal a Network Can Be
The Postal Service’s regulatory constraints come from both laws set by Congress and regulations set by the Postal Regulatory Commission. Regulatory and legal constraints have affected nearly all aspects of management decision making. Most importantly regulatory and legal constraints have slowed the ability of the Postal Service to:
• Fully integrate its delivery services up and down the communications and parcel delivery supply chain;
• Fully exploit the economic value of its intellectual, physical and labor assets;
• Adjust to changing demand for mail and parcel delivery services as quickly as the changes occur;
• Adjust its costs as quickly as its customers and suppliers can; and
• Set prices based on real market distinctions.
All of these constraints have the impact of either raising the cost of operations or lowering the potential revenue that the Postal Service can generate from its infrastructure. This forces the Postal Service to adjust its infrastructure to meet political and not business goals which results in products and an infrastructure that provides a less optimal level of service at a given price for both commercial and individual customers.
An optimal network would be free from nearly every regulatory and legal constraint that affects how the Postal Service serves its customers with two and possibly three exceptions. First, the Postal Service must deliver to every address that a sender demands. Second, The Postal Service must provide retail access to all citizens using objective criteria for measuring access. The third exception involves a very limited set of regulations of rates with a focus on single piece rates and the floor for rates for all other products.
Customer need changes over the next few decades will come faster than either Congress can pass laws or the Postal Regulatory Commission can issue opinions just as those changes have occurred faster in the past five years. Removing as many regulatory restrictions as possible would allow the Postal Service to more nimbly adjust its infrastructure as those needs change.
The optimal Postal Service infrastructure is not an ideal infrastructure. It reflects the level of service that the Postal Service can offer at set prices and costs and still generate a return sufficient to allow it to be self-sufficient. As such, the Postal Service cannot maintain an infrastructure that offer all customers exactly the level of service that they want but must try to provide an infrastructure that meets the needs of customers that it believes are most likely to require its services for the next decade or more.
The optimal Postal Service infrastructure is not static as customer needs continually change. Demand for the various services that the Postal Service offers change over time. Demand for many existing services is expected to decline over the next decade. The service and other product characteristics demanded by customers for current products will evolve as alternative communications and parcel delivery options develop.
Labor-management, capital availability, and regulatory constraints all have the effect of making the Postal Service infrastructure sub-optimal for both meeting customer needs and ensuring that the Postal Service’s financial goals are met. To the extent that that these constraints can be relaxed, the Postal Service’s infrastructure will move closer to an optimal level.