Integral Logistics Management — Operations Management and Supply Chain Management Within and Across Companies

3.1.2 Design Options for Global Distribution Networks

Intended learning outcomes: Differentiate between centralized distribution and decentralized distribution. Present features such as demand variety, need for efficient returns, degree of customer involvement in picking up. Explain design options for global distribution networks. Describe some company cases.


In a first approach and in analogy to production networks, it is possible to distinguish two fundamental types of distribution networks for delivery of a customer’s order.

In centralized distribution a customer order for a given product is fulfilled directly from the production plant or from one or a few of the manu­facturer’s central warehouses. In decentralized distribution, the given product is stored in several decentralized warehouses or with (often independent) distributors, from where the customer order is fulfilled.

The advantages of centralized distribution, from a central ware­house or the manufacturing plant to the customer, are obvious: a bigger selection of products (which, in part, can also be made per customer order, that is, “assem­ble-to-order,” “make-to-order,” or even “engineer-to-order”), a gre­ater availability, and lower total costs for inventories, plants, and handling.

The advantages of decentralized distribution are also easy to see: shorter delivery lead times and a more efficient possibility for product returns (e.g., at the retailer’s site). Besides this, transport costs are somewhat lower: first, the decentralized warehouse, which is located nearer to the customer, can be served at lower cost, e.g., using large transport units and (full) truck­loads ((F)TL); second, a customer order comprising multiple articles can be bundled into a single delivery at the decentralized warehouse. Finally, order tracking is only needed between the decentralized warehouse and the customer; if the order is raised directly in the store, tracking may even prove unnecessary. Combined with the necessary exchange of information between the manufacturer and the decentralized warehouses, the total cost for information systems is generally lower than the costs of real-time order tracking between the manufacturer and customer.

The following features for designing distribution networks have proved to be important:

  • Demand volatility, as defined in Section 3.1.1
  • Demand variety: High demand variety means that customers demand many different products. For these products the demand volatility is mostly high as well.
  • Value density, as defined in Section 3.1.1
  • Customer tolerance time, as defined in Section 1.1.2. In global distribution, delivery lead time also includes the time for customs procedures, which can disadvantage centralized distribution.

These features are highly correlated: In general, centralized distribu­tion is advantageous for products with high value density, high demand variety and volatility, and for high customer tolerance time. If the values are the opposite, decentralized delivery will be advantageous.

Two further features for the design of distribution networks, which are not correlated with the above features, however, are:

  • Need for efficient returns via the same network: Is it important that the customer be able to return goods efficiently through the same distribution network and that the network be able to handle these returns efficiently (keyword: reverse logistics)?
  • Degree of customer involvement in picking up: To what extent are customers willing and able to picking up the product themselves?

For cases where the need for an efficient product-returns process using the same distribution network is key, and for customers both willing and able to pick up the product them­selves, various designs of decentralized distribution are advantageous.

As in the case of production networks, the two groups of features often stand in opposition to one another. There are examples of this:

  • The distribution of many rather cheap items, like heavy or bulky items (e.g., bevera­ges), fresh produce (e.g., flowers), express delivery items (e.g., medicaments) or fast moving items to the point of use in companies (e.g., C items like screws, nuts, bolts, etc.): low value density (in favor of decentralized distribution), however, rather low degree of customer involvement in picking up (in favor of centralized distribution)
  • The distribution of vehicles or on-line orders: high demand variety or demand volatility (in favor of centralized distribution), however, some degree of customer involvement in picking up, as long as the pickup site is close enough (in favor of decentralized distribution)

Again a firm must make a strategic decision, which at times differs for each product family. In addition, if a company delivers to different customer segments, it will at the same time have to use different routes or channels for distribution. The channels or distribution centers need not necessarily be owned by the manufacturer. An additional required channel usual­ly entails extra costs. In addition, existing channels may change over time. For example, postal services may expand the offering of their mail rooms to become local shops, while shopping points may also offer a reduced set of services traditionally provided by a postal service.

Based on an idea in [Chop03], the portfolio in Figure 3.1.2.1 shows, in addition to the two classical designs (centralized or decentralized distribution), two possible mixed designs. The four possible designs lie in four sectors in a two-dimensional space, spanned by the dimensions that correspond to the two (conflicting) groups of features.

Sector D1 describes centralized storage near the producer or — in the case of make-to-order — delivery directly from production, with direct shipping to the customer or his unloading point. This design option is advantageous for products with high value density and high demand volatility. With this characteristic, customers are mostly willing to tolerate some time to delivery due to the generally long transport routes. This design makes possible a large selection of products, high fill rate and relatively low costs for inventory, installation, and handling. However, transport costs are rather high, as are costs for possible returns, and there are high costs for information systems for transmitting orders from the point of sale and for order tracking during shipments. This is the classic design for distribution of investment goods (such as machines), as well as for drop shipping (i.e., direct delivery from the manufacturer to the customer of the order entering party, e.g., a wholesaler, or a spare-parts or online retailer. Here, the order entering party can avoid its carrying cost but must factor in the cost of integrating its information system with that of the manufacturer and is also unable to monitor the quality of the delivered product).

The opposite sector D4 describes decentralized storage at a retailer with customer pickup. This design option is suitable if customers are able and willing to pick up the desired products, as well as products from different manu­fac­turers. It generally offers a great amount of flexibility in terms of time for this. Beforehand, customers must also do the order pick-up themselves. This design option is transparent and requires rather simple information systems for tracking orders and delivery, and it also allows returns of products or packaging material. However, as stressed in the figure, it requires an adequate retail network. Here see Section 3.1.3, particularly Figure 3.1.3.2 and the respective examples.

Fig. 3.1.2.1        Features of and design options for distribution networks.

The intermediate sector D2 describes decentralized storage in the distribu­tion center of a wholesaler or retailer with shipping to the customer or his unloading point. This is the most convenient design option for customers. But it requires rather low demand volatility as well as the customer’s pre­sence at the unloading point. Other­wise, it entails high transport costs, also owing to irregular delivery tours. The problem of optimum routing and scheduling often occurs anyway, especially with respect to efficiency for last mile deliveryor same day delivery. Using lockboxes for unloading, similar to post office boxes, the customers‘ presence may not be necessary in every case. Storage by the wholesaler copes with smaller stock levels than a corresponding retailer network, but generally does not permit same day delivery. Where the unloading point is set up to handle returns, these can also be processed by this system (e.g., the return of emp­ty bottles on a milk run). However, normally returns must take place via a different network (e.g., via the postal service’s network). This solution is suitable for the delivery of heavy articles such as beverages, of fresh produce such as flowers, of express-delivery items such as medicines, or of fast moving items, e.g., the distribution of ranges of C item goods (screws, nuts, bolts, etc.) to the point of use in firms. In the latter case, stocks may be managed by the customer (vendor-owned inventories, VOI).

The intermediate sector D3 describes delivery directly from production, or centralized storage in the distribution center of the wholesaler, with shipping to the pickup site. This design option can be selected if customers are willing and able to pick up the goods and thus profit from considerably lower transport costs. Examples include the shipping of vehicles or online orders (click and collect). But this places higher demands on the accompa­nying information systems than in the case of shipping to the customer. If the pickup site entails high costs, this solution will, in addi­tion, tend not to be cost effective. For this reason, pickup sites should be able to be combi­ned with existing distribution centers for other products or service centers (e.g., a car showroom or a supermarket chain such as Coop or 7eleven). In this case, they are also suitable for product returns or return of packaging material. Storage in the distribution center of the wholesaler reduces deli­very lead times. But it also either reduces product selection and availability or increases inventory costs. The costs anyhow increase for installation and handling owing to the costs of the distribution center. Distribution centers, including those located directly in the factory, may also act as a pickup site. An example of this is the pickup of cars from the plant.

Company Cases: When characteristic features change, it is appropriate to consi­der changing the distribution network design. In the example of Holcim mentioned in Section 3.1.1, not only a more centralized pro­duction network design but also a more centralized distribution network design became possible. Still, decentralized storage for basic demand of common products at various so-called “terminals” is part of Holcim’s distribution concept in the US Midwest (i.e., Sector D4 or D2 in Figure 3.1.2.1), right down to the Gulf of Mexico. However, for products with volati­le demand, Holcim rather uses concepts with more centra­li­zed storage (Sector D3, where the “terminals” serve as pickup sites, or even Sector D1).

At Hilti, decentralized storage in the distribution center of the wholesaler or retailer (i.e. sector D2) and subsequent delivery to the production site is executed in order to offer short delivery lead times to the customer (“last mile”). Although the inventories and respective current assets are high, the availability of the products is more important in order to satisfy the customers’ demand as fast as possible.

Further Correlations That Should Be Considered for an Integrated Determination of the Design Options: The four design options cannot be selected without giving consideration to the design of the production network. For production destined for the global market (sectors P1 and P3 in Figure 3.1.1.2), all four design options for the global distribution network come into question. In the case of decentralized storage, there is a need for a distribution network structure, possibly with multiple structure levels or echelons (see Section 3.1.3). For production destined for the local market (sectors P2 and P4 in Figure 3.1.1.2), only the design options in the sectors D2 and D4 in Figure 3.1.2.1 come into question from a global perspective, i.e., decentralized storage. From a local perspective, it is naturally possible to view storage in close proximity to a (local) manu­facturer as “central.” In such a case, all four design options can come into question, albeit only for the local distribution network.



Course section 3.1: Subsections and their intended learning outcomes

  • 3.1 Design Options for Integrated Production, Distribution, Service, and Transportation Networks

    Intended learning outcomes: Explain design options for global production networks, distribution networks, service networks, and transportation networks. Describe the network structure for decentralized distribution, and design options for retail networks. Disclose the integration of the portfolios.

  • 3.1.1 Design Options for Global Production Networks

    Intended learning outcomes: Differentiate between centralized production and decentralized production. Present features such as demand volatility, supply chain vulnerability, economies of scale, demand for consistent process quality, customer proximity, market specificity of products, value density. Explain design options for global production networks. Describe some company cases.

  • 3.1.2 Design Options for Global Distribution Networks

    Intended learning outcomes: Differentiate between centralized distribution and decentralized distribution. Present features such as demand variety, need for efficient returns, degree of customer involvement in picking up. Explain design options for global distribution networks. Describe some company cases.

  • 3.1.3 Network Structure for Decentralized Distribution, and Design Options for Retail Networks

    Intended learning outcomes: Disclose the distribution network structure and describe decision variables in its design. Present features such as available time for shopping, and simultaneously, capacity of an available means of transport of the customer, as well as the required geographical catchment area. For decentralized distribution, explain: portfolio for designing retail networks retail networks.

  • 3.1.4 Design Options for Global Service Networks

    Intended learning outcomes: Differentiate between centralized service and decentralized service. Present features such as the mobility cost ratio of the service, the degree of customer involvement in bringing and picking up the service object, as well as the need for repeated transfer of the service object. Explain design options for global service networks for services in direct contact with the object. Describe some company cases.

  • 3.1.5 Design Options for Global Transportation Networks

    Intended learning outcomes: Differentiate between direct transport and indirect transport. Present features such as size or weight of the delivery, possibility of using an existing transport network, and need for merged transport. Explain design options for global transportation networks. Describe some company cases.

  • 3.1.6 Integration of the Portfolios of Design Options

    Intended learning outcomes: Describe the interrelation between and integration of the production, transport, distribution and retail network.

Print Top Down Previous Next