FoodLegal

Designing Supply Chain Systems to Cope with Catastrophes

Pacific Asia Conference on Information Systems (PACIS)

PACIS 2009 Proceedings

Association for Information Systems Year 2009

Reeva Lederman, University of Melbourne, reeva.lederman@unimelb.edu.au
Shera Kurnia, University of Melbourne, sherahk@unimelb.edu.au
Joseph Lederman, Deakin University, Australia, joe@foodlegal.com.au

This paper is posted at AIS Electronic Library (AISeL).

http://aisel.aisnet.org/pacis2009/92

Abstract

This paper asks whether technology supported business processes currently used in supply chain management systems (SCMS) are equipped for maintaining business activity in times of crisis. It considers how SCMS accommodate the attempts of Western governments to introduce initiatives for preserving the food supply during pandemics or any other catastrophic event when there is an actual or potential impact on supply or demand. When such events occur, food manufacture and distribution is interrupted, and demand patterns are suddenly altered to become less stable and less predictable. It finds that SCMS which emphasise low buffer stock, enabled by better information sharing among parties in the supply chain through the use of IT, pose serious problems for the maintenance of the food supply at critical times. This raises the question of how supply chain processes and technologies currently employed by the food and grocery industry accommodate the supply of essential goods at times of crisis. To address this question, this research in-progress intends to model the current systems to find out what processes and technologies are most appropriate for finding an equilibrium point between the commercial interests of industry in managing inventory efficiently and government objectives in maintaining the food supply.

Keywords: Supply chain systems, supply chain technologies, supply chain systems modelling.

1 INTRODUCTION

Catastrophes during 2008, such as the tidal wave in Burma and the earthquake in China, have focused the world’s attention on how governments manage the distribution of food and fuel in situations of sudden shortfall and breakdowns in the normal distribution processes (Potera, 2008). Concern about possible pandemics such as those triggered by a human-to-human Avian influenza strain, add an extra dimension of anxiety with regard to how current food distribution systems are equipped to deal with chaotic situations and situations where the food chain needs to be protected from contagion (Fielding and Lam, 2007). In such situations there may be high rates of absenteeism from manufacturer to supermarket as a result of illness and food distribution centres may be closed or barely operating (Nieburg, 2005).

Governments world wide have introduced recent initiatives to deal with possible food and fuel shortages in an emergency. Examples include voluntary recommendations to households with regard to food storage in Australia and New Zealand and strategies of the Homeland Security Council in the US for food distribution in the event of a pandemic. All of these strategies depend on governments and the public having access to stored goods.

However, these concerns about the food supply come at a time where sophisticated information systems and technologies are being used to deliberately reduce the amount of food in storage. One of the characteristics of modern supply chain systems is reduced need for the warehousing of excess supplies. This is achieved through communications technologies which make it possible to monitor stock levels across supply chains and for the various members of the chain to respond tactically to the information provided. A prime feature of modern systems is their ability to use technology to predict demand at each point in the supply chain so that resources are not tied up in excess inventory. 

This research-in-progress paper highlights a number of potential issues and challenges that may arise during a crisis in managing food supply using modern IT-enabled supply chain systems. The aims of the overall project are:

• to examine whether technology supported business processes currently used in supply chain management systems are equipped for maintaining business activity in a time of crisis
• to identify possible ways to adapt supply chain systems enabled by various information technologies (IT) in order to provide sufficient supply of food in times of threat to the food supply.

The achieve the above aims, the Australian grocery industry is used as the context of study. A multiple case study with a number of grocery industry participants was involved in this study. There are two stages involved in this project. The first stage is to identify the current product distribution and supply chain systems in Australia. This stage has been commenced. In the second stage, detailed explorations will be undertaken, through modeling, of how different technologies or approaches could reduce the risks inherent in current systems. The overall findings of the study so far show that there is an inconsistency between the objectives of government initiatives and the objectives and capabilities of current supply chain systems. In the later stage of the research, we will identify possible ways to adapt the current supply chain systems to enable them to provide sufficient food supply in times of crisis. This may require the involvement and support of government. Thus, findings of this overall research would offer valuable contributions to both theory and practice.

The structure of the paper is as follows. The paper firstly details situations where the distribution of food and the ability to maintain adequate supply has been or could be problematic. It then examines recent government initiatives to deal with these problems and considers these initiatives in the light of current supply chain technologies and processes and their impacts. It then discusses the potential problems in cases of crisis with regard to the current trend towards managing the supply chain using approaches that emphasize minimizing or eliminating buffer stocks within the supply chain. Finally, the preliminary findings of the study are presented and discussed, followed by the plan for the next stage of the research and the conclusion.

2 FOOD SUPPLY ISSUES DURING A CRISIS

In recent years there have been a number of examples in Western countries, well serviced by large supermarket chains, where severe food distribution problems have occurred in response to natural disasters. For example, in August 2005 the city of New Orleans experienced a devastating hurricane (Hurricane Katrina). Over 80% of the city was flooded and many buildings where food supplies were produced were destroyed. Thousands of people including many food workers left the city and abandoned their employment in the aftermath (Petterson, et al., 2006). Many food shops were ransacked and supplies of food ran out quickly. The mop up continued for many months and the rebuilding still continues today.

One of the difficulties for New Orleans was that the population is predominantly poor and among those least likely to have excess food supplies at home (Loewenberg, 2008). Such populations are frequent visitors to supermarkets and food suppliers compared to wealthier populations where food purchases are made less often and in greater quantities. However, after the hurricane many shops were closed and inventory ran out quickly in those that were open. Consequently, even in households where homes were not destroyed and inhabitants not injured and able to shop, there was a lack of food. The devastation caused by Hurricane Katrina attracted more attention than other disasters before or since, even though natural disasters affecting large populations occur regularly around the world. This was largely as a result of the high concentration of news media outlets in the United States and the concern of western media for stories that affect the Western world. This attention has given rise to significant soul searching in other Western countries with regard to how they might use current systems and technologies to manage a disaster of this magnitude, while ensuring that vulnerable citizens have continuing access to essential supplies of basic food products.

Weather and environmental conditions in some countries reduce the threat of natural disaster. However, all countries are potentially subject to disasters such as pandemics or health scares such as the SARS outbreak which affected a number of Asian countries in 2002-3. During the SARS epidemic many people in countries such as Singapore stayed in their homes for fear of contagion, while in Taiwan over 1500 people were forced to stay home on home quarantine (Twu, et al., 2003). Such behaviours considerably change the usual pattern of food purchases. 

In both the United States and Australia, substantial recent efforts have been devoted to initiatives for managing an Avian flu outbreak. The US government has devoted 7.1 billion dollars for dealing with an Avian flu pandemic (Anderson, 2006) and in Australia $160 million was pledged over 2008-9 on top of a previous 133.6 million committed in the Federal Budget in 2004-5 (Lederman and Lederman, 2008). This suggests that the threat of such a pandemic is considered very real.

Such health disasters will bring with them the same conditions as were seen at the time of Katrina: large numbers of food businesses closed as a result of absenteeism, insufficient inventory in stores and reduced opportunities for replenishment and low rates of food storage among the poor. Outbreaks of illness also provide additional problems such as the possibility of quarantining of infected foods, thus removing them from available stocks. Additionally, it is anticipated that changed purchasing patterns will occur if shoppers are scared to leave the house or legally prevented as in Taiwan, because of fear of infection. When people shop they will want to purchase greater volumes of food to avoid more frequent shopping trips. Concerns about such issues have led to a number of government initiatives aimed at maintaining food supplies.

3 GOVERNMENT INITIATIVES FOR MAINTAINING THE FOOD SUPPLY

Western governments have responded in a variety of ways to the fear of food shortages as a result of natural disasters or pandemic. In most cases these responses take the form of recommendations raised in parliament or released to the public through governmental reports or white papers, rather than in the form of legislation. The idea that food storage needs to be increased at both the individual household level and the supply chain level is common to all of these responses.

The Australian Government in 2007 released an emergency pantry list recommending that all Australian Households stockpile two weeks of food supply. This list, developed by the Federal Government supported Food Industry Working Group, recommends that all Australian households have two weeks supply of non-perishable goods, contrary to common current patterns of behaviour. The official New Zealand government web-site recommends a similar course of action, also suggesting that all citizens purchase at least two weeks worth of food to cope with any emergency shortage.

In the US, calls have been made to Congress to develop Strategic Grain Reserves to deal with unpredictable disruptions in grain production. These would be similar to the Strategic Petroleum Reserve, which is an emergency fuel store of petroleum maintained by the US Department of Energy. Under the 1996 Freedom to Farm Act, the United States eliminated all its government stocks of grain and there is now considerable concern that a significant grain storage initiative ought to be undertaken.

While the thrust of various government initiatives has been on food storage, this approach is in sharp contrast to current approaches maintained by the food and grocery industry in the countries mentioned above. Food supply chain participants within the grocery industry always emphasise the need to reduce food inventory across the grocery supply chain. They do this in order to improve profitability through greater efficiencies and the implementation of supply chain strategies that require less or no inventory, enabled by the deployment of various IT strategies which result in better information sharing among the supply chain participants. However, the consequence of low inventory levels could turn into a major problem of widespread food shortages within a very short timeframe (probably less than a week) in the event of any major crisis or catastrophe such as a human influenza pandemic. In the following section we examine the literaure on risk management in supply chains and current supply chain technologies and procedures to consider how each might impact on the maintenance of food supplies in a time of crisis.

4 LITERATURE ON RISK MANAGEMENT IN SUPPLY CHAINS AND DIFFERENT GROCERY SUPPLY CHAIN SYSTEMS

Two significant studies examine the possibility of supply chain interruption as a result of strikes, terrorism or natural disasters (Craighead, et al., 2007, Kleindorfer and Saad, 2005). Both of these studies discuss in general terms the different ways of managing disruption risks. Kleindorfer and Saad (2005) separate the risks involved in co-ordinating supply and demand that occupy supply chains at all times and the risks arising from abnormal disruptions to activity. These authors propose a framework which relies on ten principles for dealing with risk. The one that stresses the importance of the type of supply chain system in place is the fifth principle which suggests that “attention needs to be given to the tradeoff between “robustness of the supply chain to disruptions and the overall efficiency of the supply chain under normal conditions”. Kleindorfer at al (2005) suggest for future research that different supply chain environments be analysed to give rise to different forms of risk mitigation although they do not analyse different types of systems in their research.

Craighead at al (2007) consider a number of supply chain design characteristics, supply chain density, supply chain complexity and node criticality. They make a number of proposals which can be summarised as follows:

• the more dense a portion of the supply chain, the more severe a disruption if it occurs in that portion 
• the more complex a portion of a supply chain, the more severe a disruption if it occurs in that portion 
• the more critical nodes in a supply chain, the more impact a disruption will have on supply.

Craighead at al (2007) propose a model which factors in density, complexity and node criticality along with opportunities in the system for warning about disruption and chances for recovery. 

While both these studies discuss in general terms different types of supply chains and their general characteristics, they do not relate their research to specific supply chain strategies and the technologies used to support them used in business today. Consequently it is worthwhile examining these different strategies more precisely both in the literature and in our own case studies.

Many different strategies are discussed in the literature and commonly implemented in industry. Due to the nature of the grocery industry that involves high volume transactions with low profit margins, the industry has been pioneering the use of information technologies to improve efficiency (Al-Sudairy and Tang 2000). Based on the use of various technologies, several of these distribution/replenishment strategy initiatives have been introduced to improve supply chain management as a result of earlier documented inefficiencies (Kurt Salmon Associates 1995; 1995/6). In particular, under the banner of the Continuous Replenishment Program (CRP), various distribution strategies have been introduced as an alternative to the traditional approach (also known as the ‘pick-and pack’ approach) in order to improve product flow within supply chains. These strategies include direct-store delivery (DSD), vendormanaged inventory (VMI), Cross-docking, Flow-through and Click and mortar. These strategies typically rely on a partnership among the supply chain members in sharing consumer demand information captured at the Point-of-Sales (POS). The key to successfully implementing such programs is complete, timely and accurate Point of Sale data, inventory level information, and a willingness and technical capability to share information between supply chain participants (Raghunathan and Teah, 2001). Capturing demand information at the POS level is facilitated by various technologies including product numbering system, barcoding systems and scanning technologies. The information sharing is facilitated by Electronic Data Interchange (EDI) employing various media including direct connections using telecommunication technologies between trading partners, the use of Value Added Networks (VANs) or the Internet. To keep track of inventory levels, systems such as Warehouse Management Systems and Inventory Management Systems are commonly used, which can also be part of an Enterprise Resource System (ERP). Significantly, those strategies that rely more on information sharing between trading partners involve more frequent deliveries and smaller quantities within each delivery, compared to the traditional pick-and-pack approach.

In general, various replenishment strategies used by the grocery industry can be distinguished based on the following characteristics:

Use of a Distribution Centre: Products supplied by manufacturers may pass through distribution centres (DC) of retailers or are directly delivered to retail stores, bypassing the DC. Strategies involving a DC can be further differentiated, depending on the existence of buffer stock at the DC. Strategies that do not involve a DC tend to have less stock in the system overall. Craighead et al (2007) consider a distribution centre a critical node because it distributes materials to many other nodes in the same chain and therefore supply chain systems that use distribution centres are particularly vulnerable.

Replenishment principle (‘push’ or ‘pull’): Under replenishment strategies which use a push principle, the supplier (manufacturer) triggers the replenishment of products and thus places orders on behalf of the customer (retailer), whereas with the pull principle, the customer triggers the replenishment.

Based on the above characteristics, Table 1 summarizes a number of distribution strategies that are commonly used within the grocery industry. For each strategy, we indicate whether the strategy involves buffer stock, uses a distribution centre and is based around push or pull approaches. Each of these strategies is discussed following the table.

The traditional distribution strategy (pick-and-pack) involves a DC with buffer stock and replenishment being triggered by customers (retailers) (Davenport and Brooks, 2004). It does not require high informational coordination between trading partners and therefore has been widely used for fast moving items. A complex IT infrastructure to manage buffer stock within a finite space and to handle daily operations within distribution centres is required for a high level of efficiency. Because of the inefficiency and high inventory level involved, in many organisations, various CRP strategies shown in Table 1 have been introduced to replace this approach.

Direct store delivery (DSD) is a strategy that is concerned with replenishing inventory at the store level. Either supplier or customer can trigger replenishment. With the advances in technology, DSD can be implemented more easily and with less cost (ECR Central 1997), with the information for reordering coming directly from customer demand. DSD is the preferred method for delivering products that are perishable, fragile, and extreme in density (do not allow efficient utilisation of trucks), require special handling and payment by regulations and have unique sales pattern (such as slow moving items with high variety and impulse).

Vendor-managed inventory (VMI) is a replenishment strategy where the supplier is given the authority and responsibility to manage the retailer’s inventory at the distribution centre level or store level (Disney and Towill 2003; Purdum 2007). Suppliers trigger the replenishment activity. With this approach, the supplier requires DC withdrawal information or POS data and the current balance of the inventory from the retailer. VMI allows the supplier to optimise their production planning, inventory level and delivery schedule (Disney and Towill 2003; Lee and Chu 2005).

A modified application of VMI, which was first introduced by Wal-Mart in the US, delays the transfer to the retailer of ownership of the inventory from the vendor until as late as possible in the period prior to the ultimate acquisition by end-consumers (Andel 1996). This approach is also known as Consignment VMI (Lee and Chu 2005). The proof of consumption obtained at a retailer’s checkout counters triggers the payment to the vendor. This approach requires vendors to have access to POS data of retailers in order to bill the retailers for the products. Under this system suppliers have control over the inventory level of the supply chain and obtain consumer demand information from the retailer (Lee and Chu 2005).

Cross-docking is a distribution strategy that involves a DC without any buffer stock where the customer initiates replenishment. It requires high information coordination between supply chain participants (;Lee and Chu 2005; Kurnia and Johnston 2001). Individual stores generate orders for replenishment and suppliers receive these orders electronically and deliver individual store orders to a DC to be sorted and collated for each store.

Flow-through, a step beyond Cross-docking, further reduces the transit time of goods at the DC. The terms Flow-through and Cross-docking have not been used consistently. For example, what is understood as Cross-docking by the Australian grocery industry is known as Flow-through by the US grocery industry and vice versa (Kurnia and Johnston 2001). With the flow-through strategy, pallets delivered by suppliers are for specific individual stores and, hence, no sorting is required at the DC. These pallets are brought directly to the dispatching areas and loaded into a retailer’s truck, ready to be delivered to stores (Vis and Roodbergen 2008). Flow-through is normally applied for products with high demand and products with highly predicted demands such as promotional products (Vis and Roodbergen 2008; Kurnia and Johnston 2001). Finally, a more recent approach, known as Click and Mortar (CAM), is typified by an online merchant in Japan to manage its retail stores. Under this approach, the end consumer orders are taken via the merchant’s website and orders are then picked up at a nearby retail store.

As Table 1 indicates, most of the recent alternative approaches do not involve a DC or do not store inventory at the DC and thus may contain less critical nodes (Craighead, et al., 2007). The ideal implementation is targeted at the store level where the POS (point of sale) data that represents the actual demand triggers the replenishment, although for the VMI approach, there is still a possibility of storing inventory at the DC. With better coordination between the suppliers and the retailers in sharing individual store demands, items do not have to be stored at the DC as they can be delivered directly to the store employing Cross-docking, Flow-through or even DSD, depending on the nature of the products.

5 RESEARCH GAP

The above literature review shows a trend toward transferring the inventory holding costs toward the supplier rather than the retailer, with the supplier bearing the risks of demand uncertainty (Lee and Chu, 2005). In return, the suppliers gain access to the actual consumer demand information captured at the POS and gain more control over the inventory level decision and management within the supply chain. Other characteristics of these recently developed distribution methods include the diminishing role of DC as a storage point (for cross-docking and flow-through) or the elimination of DC as in the case of

CAM and possibly VMI, and the trend of viewing the retailers as the infrastructure providers to sell manufacturers’ products as demonstrated in the cases of CAM and Consignment VMI. This concept can also be applied to other approaches if the transfer of inventory ownership only occurs at the retailer’s POS check out.

Both VMI and flow-through are push approaches where the manufacturer triggers the replenishment process which necessarily implies less accurate understanding of demand than if the demand was triggered by customers. Additionally, as discussed earlier, general demand responsiveness is further diminished when powerful retail chains push for goods to be delivered straight to retailers or to be stored at their own distribution centres rather than in centres not related to individual retailers where they could be more widely dispersed in an emergency once demand was known.

Cross-docking and Flow-through may have some value in times of catastrophe as using these methods products can be moved quickly to the points where they are required, provided that manufacturers are able to function and have the capacity to produce the required products. Thus, it is likely that demand can still be fulfilled despite the fact that no buffer stock is held at the retailers’ DC.

The pick-and-pack approach would probably provide the most responsive option at a time of uncertain supply and demand because there are higher levels of buffer stock (at least at from the retailers’ point of view) than in the alternative approaches. Additionally the Click and Mortar approach could be valuable in keeping customers away from stores in a situation of contagion and in making demand more fully predictable at the retail level. However, the burden of the cost of fulfilment of cancelled or uncollected orders may need to be shifted so as not to be a burden on the manufacturers and may require subsidization by governments. In devising a modified approach, it will need to be borne in mind that, in a free market economy, governments cannot easily dictate supply chain methods to businesses or demand that they take measures which would decrease their profitability in the normal (noncatastrophic) course of events without adequate compensation or subsidization from governments or cross-subsidization from others.

These recent distribution approaches all emphasise inventory level reduction as a measurement of their success. This will potentially create serious problems for governments that are concerned about having some sort of a food buffer at times of national catastrophe. Current supply chain system approaches can be seen to be working against governments trying to implement recommendations for increasing food storage overall. However, no accurate and concrete evidence is available at this stage regarding the capability of these various distribution strategies in dealing with extreme situations such as where there is a sudden increase in demand for particular products, or where production is interrupted by a high percentage of workers becoming unavailable, which can happen in the case of a national catastrophe. Moreover, where current systems automate fresh orders for inventory from point of sale based on actual consumption, the automated production systems will be severely disrupted for manufacturers, irrespective of their productive capacity, when the receiving of orders is no longer occurring in a predictable manner based on regular buying patterns.

In sum, it would be valuable to have informed knowledge, through analysis, about how the different replenishment strategies and technologies function in extraordinary situations. This knowledge would be used to evaluate the respective benefits of each approach and to inform governments about viable courses of action and the likely costs that would be borne with or without adequate compensation or subsidisation from governments or cross-subsidisation from others.

6 RESEARCH METHOD AND DESIGN

To address the gap discussed in the previous section, as described in the Introduction, the aims of the overall study are:

• to examine whether technology supported business processes currently used in supply chain management systems are equipped for maintaining business activity in a time of crisis
• to identify possible ways to adapt supply chain systems enabled by various information technologies (IT) in order to provide sufficient supply of food in times of threat to the food supply.

Specifically, the research questions are twofold:

• how suitable are the Information Technology (IT) enabled supply chain systems currently used by the food industry in many Western countries in times of catastrophe?
• what impact do these catastrophes have on the flow of information required to maintain effective supply chains?

The Australian grocery industry is considered as the context of study since the industry has been adopting and adapting various IT enabled distribution strategies and business process management strategies developed and introduced in Western countries particularly the USA. A multiple case study involving a number of organizations within the Australian grocery industry is the research method used in this study. Interviews, business documentation reviews and site visits were used as data collection techniques. The unit of analysis is the entire supply chain.

There are two stages involved in this research project. The first stage is to identify supply chain strategies used in the grocery industry today, while the second is concerned with exploring through modelling how different technologies or approaches could reduce the risks inherent in current systems. We have commenced the first stage of the research program as a means of producing data to input into the modelling stage and produced some initial results as discussed in the next section.

7 PRELIMINARY FINDINGS (STAGE ONE)

A multiple case study involving one major Australian retailer and three manufacturers was conducted to identify replenishment strategies currently used within the Australian grocery industry. Ten interviews in total were conducted with a number of managers that are involved in supply chain management. Multiple site visits and documentation reviews were conducted at all sites. Since the retailer manages their own product distribution, the distribution function of the retailer was also examined. In addition, all the participating manufacturers are the suppliers of the participating retailer. Therefore, there are five supply chains involved so far in stage 1, as shown in Figure 1.

The multiple case study shows that the retailer is currently practising the pick-and-pack, direct store delivery, cross-docking and flow through distribution strategies. Vendor-managed-inventory was trialed by the retailer and company B (manufacturer) and a number of other manufacturers, but it is no longer operational because both retailers and manufacturers think that VMI is not appropriate since it still involves maintenance of a buffer stock. Details of each current practice are provided below.

7.1 Pick and Pack

The participating retailer (Company C) has been practising the pick-and-pack approach with company B and many other manufacturers. The participating retailer’s DC with the pick-and-pack was inspected. It has been operating for almost 20 years within its 350,000 square feet. The average stock holding time is 12 days. With this amount of buffer, the distribution centre offers a high service level to the stores as product replenishment can always be fulfilled and therefore, this DC handles medium to fast moving items. A Warehouse Management System (WMS) and a computer-aided ordering system known as Reorder Inventory System (RIS) with some basic forecasting functionality are used to manage the inventory and ordering. These two systems interface with each other. On average, the handling cost per carton is comprised of 24% direct labour and 76% overhead costs, for consumable costs such as stationary, wrapping and fixed overhead costs for insurance, electricity, building, administration and infrastructure. The retailer stores place small orders to this DC on a daily basis, while orders to suppliers are generated less frequently and in a large quantity. Purchase orders generated by the RIS to company B are sent via EDI, while to other manufacturers that are not EDI capable, orders are placed via fax or phone. Upon receiving the delivery from company B, details of the orders can be retrieved from system because company B is capable of sending the Advance Shipping Notice via EDI. The participating retailer is working towards lowering the range of products handled by this approach to reduce the inventory level at the DC.

Site inspection of the participating retailer’s DC that handles cross-docking and flow-through operations was undertaken to assess the operation of cross-docking and flow-through. Cross docking is used to handle slow moving items or indent items, such as imported general merchandise due to the constraints of the sorting process at the distribution centre, which only allows each store to place orders based on a four-day roster. Products are delivered to this distribution centre by suppliers in the morning and are dispatched to individual stores in the afternoon or, at the latest, within the next 24 hours. There is no buffer stock involved. The distribution centre has been operating for about 15 years within an area of 10,000 square feet. Compared with the pick-and-pack, the cross-docking operation handles one sixth of the throughput of the pick-and-pack operation using only one thirty fifth of the floor area of the pick-and-pack operation and thus the retailer is inclined to consider increased introduction of this method. The flow-through strategy is similar to the cross-docking operation, but items delivered do not need to be sorted but ready to be delivered to the stores. This is used to handle promotional items.

The participating retailer is also using DSD with a number of suppliers including company A. A site visit to the manufacturing plant and distribution centre of company A was undertaken to examine the replenishment operations of the manufacturer for direct-store delivery. One important observation obtained from the study is that DSD requires low levels of electronic commerce technology implementation for information sharing, since the manufacturer’s representatives directly monitor the performance of the manufacturer’s products at the store level and take action accordingly. DSD also requires of manufacturer’s high warehousing efficiency to enable them to be responsive to customers’ needs and to deal efficiently with small, frequent orders. Company A chose to fully automate its warehouse for the sake of operating efficiency.

8 DISCUSSION (STAGE 1)

The case study interviews and analyses have revealed that new approaches to supply chain management have become increasingly popular due to their efficiency in reducing inventory levels and therefore costs across the supply chain. These have resulted from technological solutions that make communications within industries possible. Various systems which allow suppliers to monitor and be responsive to demand in ways not previously possible combine with the different processes discussed above to increase levels of responsiveness to demand. The merits of these approaches are crucial for the grocery industry to remain competitive despite its generally small profit margins. >

The multiple case study indicates that the participants are moving away from the traditional Pick-and-Pack (P & P) approach which involves buffer stock at the retailers’ DC level to approaches that can eliminate the buffer stock at the DC such as cross-docking and flow-through. As stated earlier, a distribution centre can be considered a critical node because of its connection to multiple nodes (Craighead 2007) so this reduced reliance on the distribution centre may have some impact on minimising disruption during a crisis.

However as previously stated the higher levels of buffer stock found in Pick and Pack may prove preferable at times of uncertain demand. Thus the trend away from Pick and Pack may not be the most appropriate for developing systems that are able to cope well with times of crisis involving demand uncertainty. To establish the significance of these trends, comparisons of these existing use patterns need to be modelled in stage 2 as well as alternatives options. Additionally, EDI capability seems central to these systems but it would be worthwhile considering what alternative technologies or patterns of use might be possible at a time of crisis if such alternatives were required.

9 PLAN FOR STAGE TWO

There has been significant work done in the past on modelling different supply chain designs and proposing optimal supply chain systems to achieve particular objectives (Beamon, 1998, Minner, 2003, Statdler and Kilger, 2005). Stage two of the research will examine how the supply chain systems analysed in stage one above can be designed or adapted to provide sufficient supply of food in times of threat by using models which consider customer responsiveness and the altered patterns of demand that occur during crises, and which focus on minimizing the possibilities of stock run-outs. Stage one found that study participants using different distribution strategies store inventories at different points in the supply chain which will affect the flexibility and responsiveness of the supply chain. This, in turn, will impact the overall cost structure and the service performance of the supply chain. What is required is a study to explore and evaluate various distribution strategies and technologies commonly used within the grocery industry and by the study participants to identify how cost and service performance can be optimized using several scenarios. The scenarios will include possible cases that will arise during catastrophes including: (1) when there is a sudden increase in demand for particular products; (2) when there are likely to be sudden or lengthy delays in buying frequency; (3) when there are changes in both production and transportation capacity and (4) when there are major changes in supply. The results of such a study will be valuable in guiding governments in how they can work in partnership with the food industry to best manage the food supply in catastrophic situations.

By comparing a number of models for supply chain analysis (Beamon, 1998, Craighead, et al., 2007, Lee and Billington, 1993, Minner, 2003, Statdler and Kilger, 2005), this paper proposes that further research is conducted to model the various supply chain systems documented in this paper using Lee and Billington’s model. This model focuses on customer responsiveness and minimizing stockout probability (Lee and Billington, 1993), rather than other models which tend to focus on reducing costs.

Lee and Billington’s model is sensitive to the type of inventories stored at each point in the supply chain not just quantity of inventory. This is important because different types of inventory are more responsive to customer demand. For example, finished goods can go straight to the customer, whereas raw materials cannot. Lee and Billington’s model also allows for decentralised control in managing the different facilities in the supply chin which may be the most feasible approach for crisis situations. 

Using Lee and Billington’s model it is possible to model the complete supply chain as a network of single sites and to measure the mean and variance on the demand on stock at each site. Thus the different processes such as VMI, cross-docking and flow-through can be modelled with consideration given to what points in the different systems cause congestion or produce opportunities to optimise the availability of finished inventory. The data for this will come from stage one. This may lead to conclusions about what parts of the supply chain a government might want to support to stimulate supply in a viable way without necessarily becoming prescriptive about what supply chain methods private companies need implement. While retail chains are not inclined to take over the tasks of government, adequate modelling of possible scenarios could indicate how this could be done and allow governments to consider how they will bear the costs.

10 CONCLUSION

The contribution of this research-in-progress paper is in raising the issue of how suitable current supply chain technologies and procedures are for satisfying critical changes in demand for food or other essential goods. The paper raises this challenging issue and outlines a method for using what is known about different supply chain systems to model optimum approaches to supply chain management for scenarios of major disruption in patterns of demand or supply capacity. First stage results of how supply chain systems in the Australian food and grocery industry currently operate are presented alongside what is already known from the literature about these systems. A second research stage that models optimum processes is underway. This modelling will result in suggestions for how inventory can be maximised to meet sudden demand and guidelines to provide essential buffers in existing systems. These suggestions and guidelines will provide a contribution to practice by offering insight to governments on how they may use current technologies and processes to devise a system for intervening to support food supplies during catastrophic times without significantly distorting the market mechanisms to which current supply chain systems respond so effectively. Through a better understanding of the workings of the various distribution strategies it will be possible to identify the challenges and issues that may evolve in different scenarios using different supply chain methods and technologies.