LOGISTICS POSITIONING AND THE DEFINITION OF THE CUSTOMER SERVICE POLICY

Defining the most appropriate policy for customer service is one of the critical factors for the success of a company, in addition to being a way to obtain sustainable competitive advantage in the long term. In general terms, companies should choose the service policy that minimizes the total logistical cost of maintaining stocks, storage and transport for a given level of service required by the market. Basically there are two possible ways to be followed. In the first, the company may adopt a quick response policy, characterized by more centralized inventories, intensive use of express transport and little dependence on sales forecasts. In the second, a policy of anticipating demand, characterized by the decentralization of inventories, located close to potential customers, and by the intensive use of consolidated shipments.

For example, as illustrated in Figure 1a, the oil industry in Brazil adopts a policy of anticipating demand, with the stock of derivatives being decentralized, from the refineries, to several primary bases that, in turn, supply other secondary bases, to meet to the fuel stations furthest from the refineries. In addition, transportation between these facilities is done through consolidated shipments in trucks or tank wagons, the shipment of which will depend on the sales forecasts made for each link. As we know, the oil distribution sector is characterized by operating with low added value products and low risk of obsolescence, in addition to having a stable demand profile, even in recessive periods. We will see later that this combination of factors minimizes the risks associated with the decentralization of inventories, in addition to allowing for the consolidation of transport, elements that make the adoption of demand anticipation policies economically viable.

On the other hand, Figure 1b shows us that companies in the information technology sector adopt a policy of rapid response, with stock centralized in a single facility, usually the factory warehouse, and their customers served through express air transport, triggered shortly after the order placement. These companies operated with high added value products and considerable risk of obsolescence, in addition to having an extremely variable demand profile, due to increasingly shorter product life cycles. This combination of factors favors the centralization of inventories and makes the adoption of rapid response policies economically viable.

The question at this point is how to choose the best service policy, that is, the one that minimizes total logistical costs for a given level of service. In this sense, the choice of service policy to be followed should not be the result of management fads or unilateral decisions by the company's top management, but should be conditioned to a prior analysis of the various factors, such as predictability of demand and added value. of products, among others, that influence its logistical positioning.

The logistical positioning refers to the set of three integrated decisions over time, which, in addition to supporting the execution and operationalization of a given customer service policy, can provide relevant subsidies for the elaboration or reformulation of the same, given that it aims to minimize total logistics costs for a given level of service. As illustrated in Figure 2, logistical positioning encompasses decisions on sizing the network of facilities, locating inventories in this network and defining the most appropriate transport policy. Network sizing involves determining the number of warehouses, their location and their mission, that is, the market regions that must be served by each warehouse. The location of stocks relates to their degree of centralization in the network, that is to say the quantities of product that must be kept in each installation. Finally, the definition of the transport policy involves, for a certain degree of centralization, the choice of the most adequate transport mode and the procedures for consolidating loads.

Depending on how these three positioning decisions are articulated, which ultimately depend on factors internal and external to the company, the lowest cost customer service policy will be adopted for a given level of service. For the sake of simplification, the discussion that follows will have a predefined network of facilities as its starting point, that is, the number, locations and service regions of each warehouse are decisions that are not easily reversed in the short term. In addition, we will be observing the movement of products from a single origin to a single final destination in the network of facilities. This is equivalent to considering only one branch or one arm of the network, as highlighted in Figure 3.

The purpose of this article is to examine in the next sections the main factors that influence the location of stocks and transport policy in a given network of facilities. On the one hand, it will be evaluated how centralization, combined with the adoption of premium transport, can economically enable the adoption of rapid response policies. On the other hand, as the decentralization and consolidation of shipments enables the adoption of demand anticipation policies. Also discussed for each of these policies are the main features, advantages, disadvantages and their cost trade-offs.

LOCATION OF STOCKS IN THE NETWORK OF FACILITIES: CENTRALIZE OR DECENTRALIZE?

Basically, there are four factors that determine a greater or lesser degree of centralization of inventories in a network of facilities: product characteristics, demand characteristics, market demand level and degree of flexibility of the manufacturing process. Other factors such as access to real-time information and the existence of economies of scale in transport also influence this decision. We will now describe the individual impact of each of these factors on the centralization decision.

The product characteristics cover the following dimensions: added value, contribution margin and degree of obsolescence. In general, as shown in Table 1, we can say that the greater the added value and the degree of obsolescence of the products, the greater the tendency towards centralization of inventories, while the greater the contribution margin, the greater the trend towards decentralization of inventories.

On the other hand, the characteristics of demand include turnover and the degree of predictability. The larger these dimensions are, the greater the propensity to decentralize inventories, basically because the risks associated with obsolescence, loss or stranding of products are minimized. Typically, products with longer life cycles and fewer substitutes have a more predictable demand profile.

• The greater the added value, the greater the propensity to centralize inventories in a single facility in order to reduce the duplication of costs associated with maintaining safety stocks in different locations.
• The greater the degree of obsolescence, the greater the propensity to centralize inventories in order to reduce the risk of stranding (non-recovery of variable costs of production and movement of materials), resulting from wrong decisions such as, for example, sending the wrong product to the wrong warehouse.
• The greater the contribution margin (difference between the price and the variable cost of a product), the greater the propensity for decentralization in order to minimize the risk of losing sales due to the lack of immediate availability of the product in stock.

The level of market demand also affects the location of inventories, characterized by two basic dimensions: delivery time and product availability. The lead time dimension strongly influences the location of inventories in the supply chain. For example, assuming 100% product availability and a given mode of transport, shorter and more consistent delivery times are more easily achieved through physical decentralization, that is, the location of inventories closer to the end customer. On the other hand, also assuming 100% product availability, contracting express transport can allow the centralization of stocks without affecting delivery times. An extreme case of the market's level of demand is the consignment of inventories to the customer, a high-cost service policy, in which a delivery period equal to zero and 100% product availability are simultaneously guaranteed.

The degree of flexibility of the manufacturing process influences the location of stocks to the extent that it is economically feasible to postpone the execution of certain steps until the customer places an order. Under certain circumstances, final mixing, assembly and packaging operations are postponed until there is a definition as to which SKUs (stock keeping units) will be sold, thus eliminating the risks associated with the uncertainty of future demand. The packaging of beer cans in packs of 6, 12 and 24 units, the final mixing of paints in construction material stores and the assembly of optional accessories in vehicles at dealerships, are concrete examples of activities that are postponed until the final configuration of demand.

The practice of postponing the execution of final operations in more flexible manufacturing processes is commonly known as manufacturing postponement. This possibility allows products in their basic form to have their inventories decentralized by installations capable of finalizing mixing, assembly and packaging operations. Decentralization becomes economically viable for two fundamental reasons:

• as stated, products in their basic form have a lower degree of demand uncertainty (for example, the aggregate demand for a brand of beer is more predictable than the total demand for packages of 6,12, 24 or XNUMX cans),
• Products in their basic form have less added value than finished products, implying lower maintenance costs for safety stocks.

Other factors relevant to the centralization decision are access to information in real time and obtaining economies of scale in transport. The first factor creates a greater propensity for centralizing inventories, due to the shortening of cycle times and the elimination of uncertainties in order processing, allowing for a more agile response to the end customer. Economies of scale in transport, in turn, create a favorable environment for the decentralization of inventories, as they make moving products through the logistics network less costly.

IMPLICATIONS OF INVENTORY CENTRALIZATION ON CUSTOMER SERVICE POLICY

Depending on the combination and interaction of the mentioned factors, it may be economically viable for a company to centralize inventories in a single warehouse, if not all, at least part of its products. Chart 2 presents some factors that contribute to the centralization of inventories. However, for a company to be able to provide adequate levels of service to its customers in a centralized environment, especially with regard to delivery times, it must be able to implement rapid response policies.

• high added value
• High degree of obsolescence
• Access to information in real time
• Low predictability of demand
• low spin
• Inflexibility in the manufacturing process
• Diseconomies of scale in transportation
• Small contribution margin

A very common practice in rapid response policies is logistical postponement, whereby the company keeps its stocks centralized until the last possible moment, immediately sending the products to customers or final consumers as soon as orders are placed. Normally, a quick response policy implies the infrequent sending of several fractional shipments in premium modes such as, for example, air. In this environment, high transportation costs more than offset the uncertainty and duplication of material handling costs associated with decentralizing inventory across multiple warehouses. Another element that makes this policy viable is information technology, which allows placing and processing orders in real time.

In several companies, highly critical spare parts are normally centralized in a single warehouse, due to their peculiar characteristics: high added value, low turnover and low predictability of demand, lack of economies of scale in transport due to fractional shipments and high demand for product availability. In the US, as illustrated in Figure 4, companies such as IBM and Xerox have centralized their stocks of spare parts to make economically viable a policy of rapid response, in which air transport and the intensive use of information technology are its main elements.

A preliminary analysis to assess the economic viability of a rapid response policy, in a simple network of facilities comprising a factory and a single advanced warehouse, consists of identifying the relevant costs for centralization. As illustrated in Figure 5, centralization leads to an increase in transport costs, due to the contracting of premium transport to cover section (c), from the factory to the final customer. This increase in transportation costs is expected to be offset by reductions in the opportunity costs of holding inventory in the chain (in transit and in the warehouse) due to a shorter delivery cycle.

If the total costs (transport and maintenance of stock in the chain) of the operation in section (c) are in fact less than the sum of the total costs of the operation in sections (a) and (b), the centralization of inventories, in parallel with the use premium transport services enable the adoption of rapid response policies. On the other hand, if this situation does not occur, a more detailed analysis should be conducted to assess whether the reductions in material handling and order processing costs resulting from the elimination of the distribution center are sufficient to offset the increase in transportation costs. .

IMPLICATIONS OF THE DECENTRALIZATION OF INVENTORIES ON THE CUSTOMER SERVICE POLICY

On the other hand, depending on the interaction between product characteristics, demand, required service levels, flexibility in the manufacturing process and other relevant factors, the physical decentralization of inventories in several warehouses may be economically viable. Decentralization normally implies a policy of anticipating demand, that is, an operational environment in which production and distribution are pushed along the various links in the network of installations until they reach the end customer. Anticipating demand proves to be quite suitable for situations in which there is little exchange of information in real time between customers and suppliers and the products have low added value. Often, investments in information technology become prohibitive and are not economically justified when compared to the added value of the products.

Generally, a demand anticipation strategy relies heavily on sales forecasts to determine and balance inventory levels at each link in the chain, a problem that does not exist when there is centralization. Another element that makes this strategy viable is the consolidation of transportation: it is expected that the additional costs of moving materials, maintaining inventories and processing orders arising from decentralization will be more than offset by the economies of scale in consolidated transportation. The next sections present a more detailed discussion on the consolidation of transport and the balancing of stock levels at each link in the chain as logistical positioning decisions that can make demand anticipation policies economically viable.

CONSOLIDATION OF TRANSPORT

There are two main ways to generate economies of scale in transportation: consolidation in time and consolidation in space. Time consolidation consists of weekly scheduling of shipments on specific days of the week. Its implementation is more suitable for situations where the daily demand for shipments is small, the delivery time is not a critical service requirement and there is a transfer of products between central warehouses and local warehouses of the same company, as shown in Figure 6.

Basically, the feasibility analysis for time consolidation is done by comparing the expected reductions in transportation costs (resulting from fewer weekly shipments from the central warehouse to the site) and the increase in safety stock levels in the warehouse location to maintain the same level of product availability (resulting from a longer average resupply time between the central warehouse and the location).

In turn, consolidation in space involves the use of warehouses to generate economies of scale in transporting part of the route from supplier to customer. Its adoption is more recommended for situations where the delivery time is a critical requirement of the service and there is the receipt of fractional shipments from different origins or the shipment to several destinations. Figure 7 illustrates the two most common operations for consolidation in space: consolidation warehouses and break-bulk warehouses.

Specifically in break-bulk warehouses, consolidated shipments are sent there from the supplier, with the aim of being divided into several smaller shipments for later shipment to different customers. This type of facility is usually located closer to the consumer market than to the supplier, in order to maximize economies of scale in consolidated transport.

DETERMINATION OF STOCK LEVELS IN EACH LINK

Having opted for the policy of anticipating demand, with the consequent decentralization of stocks, a matter of great importance is the determination and balancing of stock levels at each link. For example, if a company operates a factory warehouse, located in Rio de Janeiro, and another for break-bulk, in Recife, what should be the average inventory level in each of these facilities for a given product?

In reality, there is no solution that seamlessly optimizes inventory levels for supply chains with more than three links. One solution that many universities and consulting firms suggest is computer simulation of the flow of products through the chain's various facilities, based on the company's recent order histories. Results referring to the movement of products, stock levels, stockouts and frequency of shipments are generated and analyzed. Different service policies can be evaluated in terms of total costs and service levels based on changes in the parameters of the inventory policy and transport modes.

Finally, there are several factors that influence inventory levels at each link. In general, as shown in Chart 3, we can say that the higher the unit costs of moving and handling products and the reliability of supply, the lower the inventory levels in a given link. On the other hand, the greater the average demand, demand variability, order processing cost, economies of scale and lead-times in transporting supplies, the greater the inventory levels.

• The greater the supply reliability, the lower the stock levels due to a lower need for safety stocks.
• The higher the unit costs of moving and handling materials in a facility, the lower the inventory levels compared to other facilities that present more competitive costs.
• The greater the average demand and demand variability in the service region of a given warehouse, the greater the levels of cycle and safety stock.
• The longer the supply lead-time, the higher the stock levels due to the greater need for safety stocks.
• The greater the economies of scale in transporting supplies, the higher the inventory levels will be due to a greater quantity of products being moved from one facility to another.

CONCLUSION

This article presented the two main customer service policies: anticipation of demand and quick response. Companies should choose the lowest total logistics cost policy for a given level of service, avoiding basing this decision on managerial fads. Determining the best policy is conditioned to an analysis of several factors external and internal to the company that influence its logistical positioning: dimensioning of the network of facilities, location of stocks and transport policy.