1. Introduction
1.1 Problem indication
Eleven percent of mankind suffers from hunger, which amounts to 795 million people (FAO, 2015). Furthermore, continuing population and consumption growth will result in a higher demand for food. According to recent studies, food consumption is expected to increase by 70% by 2050 (Burch, et al., 2007). Hence, worldwide demand for food is considerable and growing rapidly. In order to meet this demand and feed the growing population food production is increasing. But, an increase in food production has a negative effect on the environment, since enhanced production results in higher release of greenhouse gases and higher environmental pollution (Godfray, et al., 2010). A more environmentally friendly way to increase food availability is to decrease the worldwide food losses. Nowadays, approximately 30% of food produced for human consumption is wasted at one of the stages in the Food Supply Chain (FSC) (Dobbs, et al. 2011; Gustavsson, Cederberg & Sonesson, 2011). Hence, considering current food waste, increasing food production is not strictly necessary because the world’s food production is more than sufficient to feed the world’s population. The Food and Agriculture Organization of the United Nations even claims that recovering half of what is lost can feed the world. Therefore, reducing unnecessary food losses is of great importance in order to increase food availability without negative impacts on the environment.
Food is wasted to a varying extent across the world and across all stages of the FSC (Lipinski, et al., 2013). These losses have a double impact on the environment, first because of food production and second because of collection and treatment of waste (Stenmarck, Hanssen, Silvennoinen, Katajajuuri, & Werge, 2011). European countries experience an annual waste of approximately 100 million tonnes of food and it is expected that without counter-measures this will increase to 120 million by 2020 (Comission, 2016). The Netherlands is a big contributor to Europe’s total food waste and is ranked third regarding total food losses. Dutch annual food waste amounts to 9.321.803 tonnes, which is 9% of total European food waste. Number two, Germany, wastes 10.387.696 tonnes/year and list leader the United Kingdom wastes 11.887.389 tonnes/year (Eurostat, 2014). However, considering the number of residents of the Netherlands, this country is ranked high on the list of total food waste. In short, these figures clearly present a problem in Europe and especially in the Netherlands.
All stages within the FSC chain contribute to the total food waste, but most food losses occur in the retail and consumer stages. These stages are responsible for 31% of total food waste across the FSC (Parfitt, Barthel, & Macnaughton, 2010). Losses in retail have grown rapidly over the past years and are expected to keep increasing (Godfray, et al., 2010). With 443.192 tonnes/year, the Netherlands takes fourth place in the ranking of Europe’s total food waste in retail. Where for example France, with a population four times as big, waste less than double the amount of food that is wasted in the Netherlands (782.339 tonnes/year) (Eurostat, 2014). Furthermore, the Netherlands is also fourth regarding food losses in the retail sector per capita, quantified as 74kg/per year, while Belgium is ranked first with an amount of 90kg/year. Summarized, food waste in the retail sector is significant in the Netherlands. Stenmarck et al. (2011) claim that these losses are avoidable and can be prevented. Therefore, this thesis concentrates on the causes and possible solutions for unnecessary food losses in the retail sector in the Netherlands. However, this research only focuses on supermarkets in the Netherlands, because supermarkets are responsible for 80% of the turnover in the retail sector in the Netherlands (Detailhandel.info, 2016).
Unnecessary disposal of food is mainly caused by obsolescence of products, meaning that products become unsellable before they reach the market. Products become unsellable based on expiry of their expiration date or due to damage (Gustavsson et al. 2011). Unsellable products cannot be offered in stores, resulting in disposal of products (Kantor, Lipton, Manchester, & Oliveira, 1997; Stenmarck, et al. 2011). In this study unsellable products are defined as products which passed the ‘best before date’ and fruits, vegetables and bread with a lack of freshness (Stenmarck, et al. 2011). However, unsellable products are just symptoms, the real reason for unnecessary food losses is inaccuracy of the inventory systems (Thankappan, 2011). Inventory information inaccuracy is a significant problem for supermarkets that use automated inventory management systems (DeHoratius, Mersereau, & Schrage, 2005). Empirical research at 37 retail stores shows that 65% of inventory records do not match physical inventory (DeHoratius & Raman, 2004). Furthermore, Kang & Gershwin (2004) report that even the best performing retailer has a considerable inaccuracy of 30%. Supermarkets with inaccurate inventory systems experience additional costs caused by uncertainty about the level and location of their inventory. Subsequently, inventory information inaccuracy results in inefficient operation of automated replenishment and demand forecasting systems of these supermarkets. The systems will order when ordering is unnecessary and do not order when they should, resulting in excess inventory or loss of sales because of stock-outs (Raman, DeHoratius, & Ton, 2001). According to Stenmarck et al. (2011), excess inventory is amplified by customers’ desire. For example, customers expect full shelves, reject products with a short expiration date and generate fluctuating demand. These aspects create challenges for inventory control of supermarkets, because to fulfil customer demand and ensure full shelves, supermarkets often order more than necessary, resulting in excess stock which eventually causes unnecessary food losses. Hence, inaccuracy of inventory systems causes unnecessary food losses in Dutch supermarkets.
Food losses in retail (and in general) are an issue worldwide as well as in the Netherlands. Therefore, this research concentrates on a potential solution for the inaccurate inventory systems of Dutch supermarkets. According to Kang & Gershwin (2004); Atali et al. (2006) & Heese (2007), RFID (Radio Frequency Identification) technology is a solution for inventory inaccuracy. This technology could improve inventory accuracy by increasing visibility, resulting in better inventory alignment. The first passive RFID system was developed during World War II. The technology was used to detect approaching planes while they were still far away. The first commercial application was performed in 1987 and ever since RFID evolved into one of the best-known supply chain technologies (Landt, 2005). RFID can be seen as a successor of barcode technology in the Supply chain (SC) and has been proven to reduce information gaps in the SC, especially in retail (Angeles, 2005). Reduction of information gaps results in prevention of stock-outs and avoidance of excess inventory. For example, Wal-Mart has performed a successful trial of RFID technology in January 2005. The American multinational retail corporation implemented an in-store RFID tracking system with the purpose of reducing the number of stock-outs. The implementation of RFID resulted in a 16% decrease of the out-of-stock rate (Wal-mart, 2005).
The issues contributing to excess inventory in retail can be reduced by the use of RFID. The technology allows retailers and manufacturers to identify the exact quantity and location of their inventory without conducting time-consuming inventory checks and the technology provides real-time visibility (Michael & McCathie, 2005; Kok, Donselaar, & Woensel, 2006). Full visibility enables supermarkets to improve their inventory information accuracy and results in faster response to customer demand and market trends. Furthermore, it improves the ability to have the right product in the right place at the right time. Most importantly, the technology reduces inventory levels. In short, RFID is a potential solution to reduce the bullwhip effect and improve inventory management performance. The technology facilitates real-time inventory information, resulting in optimal inventory levels and replenishment orders, which contribute to the reduction of unnecessary food losses of supermarkets.
1.2 Problem statement and Research questions
1.2.1 Problem statement
Based on the problem indication the following problem statement is derived:
“To what extent could RFID contribute to the decrease of food losses of supermarkets in the Netherlands?’’
1.2.2 Research questions
Theoretical research questions:
1. What are the causes of inaccurate inventory systems in supermarkets?
This question is answered by conducting a review of the literature to asses to what extent inaccurate inventory systems are the cause of unnecessary food losses in Dutch supermarkets. Furthermore, answering this research question will provide insight in the causes of inaccurate inventory systems in supermarkets. The identified causes for inaccurate inventory information are described in detail and the effect is presented.
2. What are the characteristics and benefits of RFID?
Answering this research question provides the benefits and drawbacks of the RFID technology. In addition, the characteristics of RFID and its predecessor, the barcode, will be compared.
Empirical research questions:
3. What are the RFID project requirements and costs for the Dutch supermarkets?
This research question concentrates on the costs and requirements of the implementation and use of RFID in Dutch supermarkets. The implementation and periodic costs of RFID will be examined and the requirements for the use and implementation of RFID in the Dutch supermarkets will be identified.
4. How can the implementation of RFID influence the amount of unnecessary food losses in supermarkets in the Netherlands?
The last research question explores the effect of an RFID implementation on avoidable food losses. First, the current amount of food losses will be quantified, shrinkages figures of five Dutch supermarkets will be presented and the impact of RFID will be applied to these numbers. Subsequently, the influence of RFID will be expressed in terms of reduction of unnecessary food losses (€). Finally, it will be concluded whether RFID is an appropriate solution for the Dutch supermarkets.
2. What is RFID?
Radio frequency identification (RFID) is an identification technology that uses radio waves to transfer data between tags and readers (Landt, 2005). Basic principle of RFID is that objects, animals or persons which contain a tag can be identified and followed in real-time without requiring any line of sight between tags and readers (Ward & Kranenburg, 2006). In simple language tags say: ‘’Hello, here I am and this is my identification’’. Readers receive the information and transmit it to the host where it is processed. RFID systems consist of readers (interrogator), tags (transponders) and a host (controller), see figure 1.
Figure 1: RFID system components Source: Glasser et al. (2007)
2.1 Operation of a RFID system
RFID is based on a simple concept: two elements a “Tag” and a “Reader” communicate with each other via radio transmission. In simplified language, readers interrogate tags and receive an answer when tags are located in read range of readers. Magnitude of this read range depends on the electromagnetic field created by the coiled antennas of readers and tags. As soon as tags enter the electromagnetic field, they receive a signal from readers (Roberts, 2006). Thereafter, tags will decode received signals and broadcast a reply signal consisting of an Electronic Product Code (EPC), which is a unique number that identifies a specific item (Angeles, 2005). Readers capture the reply signal that reflects tags identification (EPC) and data content. Subsequently, readers decode returning signals and transmit data to the host that process the information by using RFID software (Weinstein, 2005). Hence, data of tags can be read and processed from a distance, simultaneously, automatically and fast.
2.2 Tags
Basic function of tags is to store data and transmit data to readers. Tags are comprised of an integrated circuit (microchip) attached to an antenna that has been attached onto a mount which is often a paper substrate. The microchip is pre-programmed with a tag identifier (EPC) and includes a memory bank to store item’s unique data. Furthermore, tag’s antenna collects energy and channels it to the microchip to turn it on. Generally, the larger the tag antenna’s area, the more energy will be received and channelled toward the tag’s microchip, and thus the further the read range (Tesoriero, Gallud, Lozano, & Penichet, 2008). Additionally, tags can be classified depending on their source of electrical power, their memory and used frequency band. Based on their electrical power source, tags can be distinguished into “Passive tags” and “Active tags”. Regarding data storage capability, tags can be classified into “Read-Only” and “Read-Write” tags. Finally, tags can be distinguished based on their frequency band in Low frequency (LF), High Frequency (HF), Ultra High Frequency (UHF) and Microwave Frequency. All mentioned classifications are described in detail in Appendix 1.
Figure 2: Passive RFID tag Source: Roberts (2006)
2.3 Readers
Readers (interrogators) act as a bridge between tags and host, because readers transmit and receive information from and to tags and host. Information from tags is captured and forwarded to the host by means of reader’s attached antennas. When passive tags are within the read range of readers, tag’s antenna absorbs energy from reader’s electromagnetic field. Absorbed energy is converted into electricity that can power up tag’s microchip, enabling tags to reflect back signals from readers. Next, antennas of readers emit radio waves into their operation field, antennas of tags receive this signal and reflect back to the readers. Subsequently captured data is transmitted from readers to the host for processing (Want, 2006; Roberts, 2006).
Readers are distinguished by their function and frequency range and have the following basic functions: read data content of RFID tags, write data to a tag (in case of read-write tags), transmit data to and from the host and power passive tags (Hunt, Puglia, & Puglia, 2007). Hence, besides tags, readers can also be divided into “Read-Only” and “Read-Write” readers. Read-only readers can only query or read information from tags, while read-write readers are capable to read and write information on tags. In addition, frequency bands are specified for each type of reader. For instance, high frequency readers and ultra-high frequency tags are not able to communicate with each other because of different frequency bands (Ward & Kranenburg, 2006). Furthermore, there are three different styles of readers. Selection of most appropriate type of reader depends on the selection criteria described in section 2.5. Readers are available in the following forms: fixed, mobile and desktop readers, see figure 3. Fixed readers are used to create a fixed read field for automatic reading. This type of reader is most commonly used to read tags when they enter locations, such as stockrooms. In order to achieve 100% readability, fixed readers contain two to eight antennas. Mobile readers are the same as mobile computers, such as smartphones and PDA’s, but with addition of an RFID antenna. These type of readers are used for manually reading of tags on the move. Finally, desktop readers are used for applications which need readers right next to computers for easy input. Desktop readers are always connected to computers which contain RFID software (Hunt, Puglia, & Puglia, 2007).
2.4 Host
Hosts (controllers) are the brain of every RFID system, because hosts manage the flow of data between readers and tags. Furthermore, hosts enable several monitoring options which are used to keep track of all data and update it when necessary. Commands from hosts are converted into radio waves by readers, tags respond to the broadcasted radio waves. This reply signal is received by readers and processed by the host. Data acquired by readers will be processed by the host into useful information, resulting in registration of in- and outgoing goods, track movements and verification of identities (Feldhofer, Dominikus, & Wolkerstorfer, 2004).
2.5 Conclusion
There is no perfect tag or reader for all RFID applications, since it is the application that defines the specifications. Some tags might be optimized for a particular frequency band, while others might be tuned for good performance when attached to materials that may normally not work well for wireless communication (liquids and metals). Supermarkets aim for tags which enable identification of incoming and outgoing goods, tracking of items and monitoring of expiration dates. Selection criteria for best suited type of tags and readers are presented below:
Desired minimum “read range” of tags/readers. Higher-frequency tags can be read from greater distances but they require a higher energy output from readers.
Country where the application will be implemented, because allowed frequency band depends on regulations of relevant country.
Desired size of the tags, are there limitations concerning the dimensions of tags regarding the method of use?
Scan and data transfer rate, what is the speed at which some readers have to be able to read all passing tags. For example, how fast will items move through portal and gate readers?
Costs: is it an economical choice that fits into the investment budget?
Nevertheless, several tagging technologies are needed for RFID systems to realize its full potential. Cost savings provided by passive tags makes RFID implementations possible at a much lower price than it would be with the use of active tags alone. On the contrary, active tags add functionalities which are not possible to perform with passive tags. Ultimately, RFID application determines what type of tags and readers will be used.
3. Theoretical framework
This chapter introduces and describes the literature used in this thesis. First, operation of the inventory management system of supermarkets is explained. Next, the importance of inventory information is analysed and the causers of inaccurate inventory information are identified and explained. Third, benefits and drawbacks of RFID deployment in supermarkets are presented. Finally, RFID technology will be compared with its predecessor the barcode.
3.1 Inventory management system of supermarkets
Most commonly, inventory management of supermarkets is performed by Automatic Replenishment Systems (ARS), which track the number of products in the store and stockroom and place automatically an order to the supplier when the level of inventory is below predetermined reorder level. Replenishment of supermarkets is based on product usage and stock level information and is triggered by actual needs of customers (Stank, Daugherty, & Autry, 1999). Purpose of ARS is to balance product availability with unnecessarily high stock levels, resulting in more efficient stocking, more space productivity due to fewer stock-outs and less stockroom space for safety stocks.
Starting value for the operation of ARS is the initial stock information. Subsequently, ARS is based on two information flows: information of incoming (ordered) and outgoing (sold) goods. Information of incoming and outgoing goods is used to monitor inventory of a supermarket continuously. Supermarkets can track the level of stocks on individual item-level by using the overview of ordered items and the Point of Sales (POS) data, which reflect actual sales figures (Raman, DeHoratius, & Ton, 2001). Ordered quantities are determined by the ARS and manual operations (adjustment of standard reorder levels) of a supermarket manager in case of sales promotions. Information about outgoing goods is obtained after customers complete their payment at one of the POS options – a cash desk or a self-checkout machine. Products scanned by the cashier or customer himself are automatically recorded and registered as outgoing products (Atalı, Lee, & Özer, 2006). Hence, supermarkets determine their inventory based on ordered goods and sold goods. The subtraction of both of these values results in the current inventory levels.
Generally, ARS determines which goods the supermarket orders, because replenishment of inventory is triggered by shelf- and stockroom levels. Whenever, inventory levels fall to a predetermined reorder level a replenishment order will be automatically transmitted to the supplier of the supermarket (Kang & Gershwin, 2004). Thus, in case insufficient stock levels are detected – caused by stock levels and ordered quantities which are lower than the predetermined reorder level – replenishment orders are automatically sent. Reorder levels are set in a way that there is still enough inventory to meet the customers’ demand until the next order comes in. The reorder level has to be well aligned with the customer demand, because if a reorder level being set too high it results in an unneeded high inventory. An understated level, on the other hand, will eventually result in stock-outs.
Furthermore, supermarkets act on two events in order to align physical inventory and recorded inventory. First one is a manual periodic inventory audit with Personal Digital Assistants (PDA), performed once a week after closing time. Second one is a signal given by the customers or employees when a product is out of stock. Weekly inventory audits are preferred over out of stock signals, because empty shelves influence the customer satisfaction, resulting in lower sales. After identification of inventory discrepancies by means of conducting audits or receiving out of stock signals, recorded inventory levels are corrected manually in ARS. Furthermore, Backes (1980) and Flores & Whybark (1987) argued that errors accumulate in the inventory records until an inventory audit is performed. In addition, Atali et al. (2008) claim that the longer errors occur in ARS the more they increase the risk of stock-outs and excess inventory. Hence, it is essential to perform inventory audits in order to detect and correct discrepancies. Besides inventory audits, a manual shelf-life control is performed daily in order to check the expiration date of perishable goods. Products that exceed their shelf life are removed from the shelves, recorded as unsellable and disposed (Broekmeulen & van Donselaar, 2009).
3.2 Inventory information inaccuracy
According to Rinehart (1960), information inaccuracy issues are an obstacle for optimal operation of inventory management systems, because inventory information inaccuracy leads to inventory discrepancy. Therefore, successful operation of an inventory management system of supermarkets depends on the accuracy of inventory information, since the optimal order ability is influenced by the accuracy of inventory information. Incorrect inventory information can lead to improper operation of ARS of supermarkets, resulting in unneeded high inventory levels or poor customer services due to stock-outs (DeHoratius, Mersereau, & Schrage, 2008). Although, supermarkets make a crucial assumption regarding the inventory information, by supposing that the level of the physical and recorded inventory is equal. However, DeHoratius & Raman (2004) reported that 65% of the inventory records are not matching the actual physical inventory. Furthermore, Raman et al. (2001) claim that retailers have numerous inventory information errors, even though they use an ARS system. Moreover, they examined that inventory inaccuracies could reduce the retailers’ profits by 10%, due to higher inventory costs and lost sales. A further research of DeHoratius et al. (2008) claims that 3.4% of products listed on the inventory records were not available in supermarkets. In addition, Kang & Gershwin (2004) found that even the best performing retailer faces a considerable inaccuracy of 30%. Although supermarkets assume to be accurate, the conclusion can be drawn that this is incorrect, meaning that in many cases the ARS of supermarkets determines its operations based on erroneous information. Hence, if inventory information transmitted to the ARS is incorrect and if the periodic controls do not account for inventory discrepancy, the inventory management system of the supermarkets will be inaccurate.
Inaccuracies regarding inventory information affect the replenishment orders, because inventory information is used to support quick decisions, associated with automatic replenishment and is subsequently used to trigger re-shipments. For instance, incorrect inventory information generates replenishment orders unnecessarily, not at all, or with the wrong quantities, resulting in stock-outs or high level of unnecessary inventory.
Figure 4: Cause effect diagram Source: Author
3.2.1 Causes of inventory information inaccuracy
Following the academic paper of Iglehart & Morey (1972), which analysed inventory systems with imperfect information, numerous academic papers were published discussing the impact of inventory inaccuracy and its causes (Atalı, Lee, & Özer, 2006; DeHoratius & Raman, 2004; Lee & Özer, 2007). According to Lee & Özer (2007), inaccurate inventory information is the result of theft and unsellable products, misplacements of products and transactional errors. Similarly, DeHoratius et al. (2008) claim that information inaccuracy is a consequence of replenishment errors, employee theft, customer shoplifting, improper handling, incorrect inventory audits, and incorrect recording of sales. The distinguished factors from literature are classified into three groups and used in order to describe the sources of inventory information inaccuracy:
Classifying presented causes allows to examine the impact of RFID on each source separately, which is essential since some causes could lead to permanent inventory shrinkage while others could lead to temporary inventory shrinkage. Furthermore, each classified cause influences recorded or physical inventory in a different way: some causes will increase inventory while other will deplete inventory. In short, inventory information becomes inaccurate due to the mentioned factors that are described individually in this paragraph.
3.2.1.1 Transaction errors
Transaction errors were introduced in inventory management by Iglehart & Morey (1972), and are related to incorrect scanning or identification of items (Raman, 2000).
These errors only affect inventory records and leave physical inventory unchanged, meaning transaction errors result in inaccurate inventory records (Kok, Donselaar, & Woensel, 2006). Inventory records will not reflect the physical stock precisely, resulting in inaccurate inventory information (Atalı, Lee, & Özer, 2006). In general, transaction errors are mainly caused by human errors and occur at the inbound side, the shop floor and the outbound side of supermarkets. According to Kok et al. (2006), transaction errors are usually caused by not scanning a product, incorrect manual registrations or multiple scanning of items. Firstly, not scanning a product is caused by a lack of precision of a store employee or customer. Secondly, incorrect manual registration includes incorrect product identification and typing errors, caused by improper selection of quantities or categories. Finally, multiple scanning of items occurs if products are scanned multiple times instead of only once.
The inbound side of supermarkets is specified as the place where the incoming goods are received. Next, the shop floor is defined as the actual part of the supermarket that is accessible for customers and the outbound side exists of all POS. Supermarkets receive daily shipments from distribution centres (DC), these incoming shipments have to be registered and updated into the ARS. However, precise measurements are never performed at the inbound side during receiving of goods (DeHoratius & Raman, 2004). When the ordered shipment arrives, a store employee will assist the truck driver unloading the truck from the DC. The store employee counts the amount of delivered containers with goods and checks the shipment on accuracy. However, only the sum of delivered containers is checked, the amount of actual incoming items is unknown (Waller, Nachtmann, & Hunter, 2006). Hence, supermarkets assume that the ordered quantity is equal to the incoming physical quantity, which in case of incorrect shipments leads to inventory information inaccuracy (Ton & Raman, 2010).
Furthermore, transaction errors on the shop floor may arise during two events: registration of unsellable products or inventory audits, for both events store employees make use of a PDA (Lee & Özer, 2007). Whenever products are damaged or have expired their shelf-life they will be disposed. Before disposing, products must be recorded in order to prevent unknown stock loss and inventory discrepancy. Nevertheless, transaction errors can occur during registration of unsellable items due to incorrect or multiple scanning. For example, incorrect manual registration can be a result of typing mistakes, leading to higher or lower reductions of recorded inventory than intended. In addition, when store employees cannot identify the discarded product correctly, other products are selected on the PDA or the disposed products are not scanned at all. Hence, transaction errors cause unknown stock losses or more/less than necessary recorded losses, resulting in inventory discrepancy.
Next, according to Atali et al. (2006), inventory discrepancies would not be discovered and identified without inventory audits. Furthermore, Iglehart & Morey (1972) claim that inventory audits completely eliminate inventory discrepancies. However, several transaction errors may happen during inventory audits, such as incorrect identification, counting and scanning errors. Incorrect identification is a consequence of the enormous volume of products handled in supermarkets; thousands of different products, and tens of thousands of items come and go in a supermarket during a single day. On account of the large variety of products, several differentiation mistakes occur, leading to incorrect registration during inventory audits. Similarly, products can be scanned multiple times or not even be scanned during inventory audits. Finally, manual errors may occur during the inventory audits, such as counting or typing errors. Summarized, transaction errors which occur during inventory audits cause transmitting of inaccurate inventory information to the ARS.
Lastly, at the outbound side, transaction errors occur as a result of human errors of customers or cashiers. Both, the cash desks and the self-checkout machines are vulnerable for human errors, because most actions have to be performed manually (DeHoratius, Mersereau, & Schrage, 2008). For instance, scanning products multiple times results in a higher reduction of recorded inventory of the particular product than intended. Furthermore, most common transaction error at the outbound side is the processing of products as others. For example, when a customer buys multiple flavours of a product which have all the same price and roughly the same packaging, such as Cup-a-Soup. The cashier or the customer scans one unit and hits the number key to multiple the units of the same product, resulting in a more than necessary reduction of recorded inventory of the scanned product. Furthermore, the physical inventory of the product will be higher than recorded inventory, while other non-scanned products have a lower physical than recorded inventory. Hence, scanning one item multiple times could result in a lower recorded inventory for the scanned item, while it results in a higher recorded inventory for the other non-scanned products. Finally, the error of not scanning the product at all results in a loss of sales and a lower physical inventory than recorded inventory.
3.2.1.2 Shrinkage
Supermarkets suppose they know exactly what their inventory level is and assume that this information is reliable. However, inventory records are only rough estimations of the physical inventory. The ARS of supermarkets does not take into account the “disappearing” inventory caused by shrinkage. In order to correct the discrepancy between the recorded and physical stock, corrective actions are performed by means of time-consuming inventory audits.
In consequence of the necessary costly inventory audits and loss of sales, inventory shrinkage is a significant problem for supermarkets, because it has a major impact upon the profitability – which is crucial in this relatively low margin sector (Kalyanam, Lal, & Wolfram, 2010). Furthermore, Atalı et al. (2008) show that shrinkage accounted for 2-4% of total sales in the retail industry of Europe. Moreover, indirect losses, such as losing customers due unexpected stock-outs, are even fifteen times higher than direct loss of inventory (Lee & Özer, 2007).
Small shrinkage rates already influence the ARS, because physical inventory level will be lower than recorded inventory level. For instance, shrinkage may cause stock-outs before the inventory reorder level is reached. Hence, ARS will not be triggered until a manual periodic control is performed, which leads to even more dissatisfied customers. Shrinkage in supermarkets is a consequence of theft, unsellable products and incorrect deliveries. Furthermore, it affects physical inventory of supermarkets, but leaves inventory records unchanged. In a nutshell, shrinkage only influences the physical inventory and can only be detected by conducting periodical audits.
Dutch supermarkets experienced a loss of income caused by shrinkage of almost 1 billion euro in 2015 (Maarse, 2015). Theft caused €420 million of revenue loss, which can be divided into shoplifting (47%), dishonest employee thefts (30%), administrative errors (16%) and finally vendor or supplier fraud (7%) (Cube, 2014; FAD, 2014). According to CBS (2014), the number of registered shoplifting cases amounts to 43.395. However, shrinkage caused by thefts is rarely registered by supermarkets, since the outbound products are not recorded at the POS of the supermarkets. These losses are not detected and thus not updated into the ARS, resulting in unknown stock losses. Second cause of revenue losses is attributable to unsellable goods and amounts to €580 million. Unsellable products are disposed and thus not available for customers, leading to a lower physical than recorded inventory. However, shrinkage caused by unsellable products has a lower impact on incorrect inventory information than thefts, because in contrast with thefts, stock losses due to unsellable products are not always unknown. As mentioned before, unsellable products are before disposal recorded as shrinkage, these recorded losses are subtracted from the recorded inventory, resulting in an up-to-date recorded inventory. However, shrinkage of unsellable products could cause unknown stock losses due to lack of correct registration (Kok, Donselaar, & Woensel, 2006). In short, unknown stock losses caused by thefts or improper handling of unsellable products result in inventory discrepancy.
3.2.1.3 Misplacement
Nowadays, supermarkets offer an enormous variety of products. Such a variety is associated with an increase of misplacements in supermarkets, because a larger assortment increases complexity and confusion at supermarkets (Ton & Raman, 2010).
Misplacements can be caused by both customers and store employees. Customers progress through a sequence of decisions while they are shopping. First, customers have a desire or interest and select a product. Subsequently, the customer might face a hesitation whether to keep or discard (reject) the selected product. When customers reject a certain product they will either return it to the correct specified location or leave the product at another incorrect location (Gnepa, 1996). Furthermore, store employees can cause misplacement by storing products at the wrong shelves or locations that are not accessible for customers, such as the stockroom (Fleisch & Tellkamp, 2005). Hence, misplacement is caused by not returning or storing products on the correct location by either customers or store employees.
Misplaced products influence the supermarket performance in three ways. First, misplacement can lead to a lower sales volume. When customers are not able to find desired products, they are not able to buy the products, resulting in lower sales. Second, misplacements cause a lower labour productivity, because store employees have to provide additional assistance to customers that are unable to find their desired products. Third, misplacements affect the ability of the ARS to monitor the actual inventory levels, because misplaced products result in temporary inventory information errors that lead to inventory discrepancy (Ton & Raman, 2010). Hence, misplacements reduce recorded inventory but leaves physical inventory unchanged, products are not lost, only misplaced and thus unavailable to satisfy customer demand. Consequently, physical inventory is correct but partially not stored at the right location, resulting in incorrect inventory records. Unlike shrinkage, misplaced items can be returned to inventory after a periodical audit, because misplaced products will eventually be found and corrected (Fleisch & Tellkamp, 2005). However, if products are returned after a period of time it is reasonable to believe that the recorded inventory does not match the new physical inventory (including returned products), resulting in a higher replenishment order than necessary (Atalı, Lee, & Özer, 2006).
3.2.2 Conclusion
As discussed above, transmission and use of inaccurate inventory information result in an inefficient inventory management system, because inaccurate inventory information influences the optimal order ability negatively, resulting in a higher or lower inventory than necessary. A higher inventory results in excess inventory, which eventually can results in unnecessary food losses. In contrast, a lower inventory than necessary can cause stock-outs. According to Prater et al. (2005), stock-outs cause a revenue loss of 3%. Nevertheless, this thesis concentrates on determining the impact of RFID on avoidable food losses, thus the research is focused on the causes and consequences of excess inventory instead of stock-outs. Causers of discussed inventory information inaccuracy that eventually lead to excess inventory are transaction errors and misplacements. Shrinkage influences inventory information accuracy, however shrinkage results in stock-outs instead of excess inventory, see table 1.
Causes of inventory
information accuracy Recorded inventory Physical inventory Duration of impact on inventory Stock-outs Excess inventory
Transaction error Higher or lower Unchanged Temporary Yes Yes
Shrinkage Unchanged Lower Permanent Yes No
Misplacement Lower Unchanged Temporary Yes Yes
Table 1: Impact causes inventory information inaccuracy Source: Author, based on: Atali et al. (2006), Lee & Özer (2007) & Fleisch & Tellkamp (2005).
In short, the inventory accuracy problem of supermarkets is caused by inventory discrepancy, which is caused by transaction errors, shrinkage and misplacements. Subsequently, inventory discrepancy leads to incorrect inventory information, which is essential for the proper functioning of the ARS of supermarkets. All mentioned causes of inventory discrepancy can lead to stock-outs due to ordering less than necessary. However, excess inventory can only be caused by inaccurate inventory information due to transaction errors or misplacements.
Figure 5: Cause effect Diagram of unnecessary food losses Source: Author
3.3 Benefits and drawbacks of RFID
Research on the benefits of RFID have mainly focused on the entire supply chain. Studies inferred from the advantages of RFID in supermarkets are limited. Furthermore, researchers mainly focus on the impact of RFID implementation on the entire supply chain over the specific impact on supermarkets. In order to fill the literature gap, this paragraph will discuss the benefits and drawbacks of RFID implementation in supermarkets by specifying reported supply chain advantages. The impact of RFID implementation on supermarkets will be explored in detail later on in this thesis. Finally, this section ends with the comparison of the technical advantages of the RFID technology and the traditional barcode technology.
3.3.1 Benefits of RFID implementation
Supermarkets determine the products offered and quantities stocked. In order to achieve inventory control, supermarkets have to make trade-offs between buffer inventories to prevent stock-outs and higher level of shrinkage or lower inventory levels in order to reduce unnecessary food losses and higher possibility on stock-outs (Koh, Kim, & Kim, 2006). Quick response technologies, such as the barcode, helped supermarkets to improve product tracking, reduce labour costs and speed up product replenishment. The ability to speed up the replenishment process allowed supermarkets to hold reasonably low inventory and reduce overstocking (Attaran, 2007). However, visibility generated by barcode technology is limited compared with the full visibility facilitated by its successor the RFID technology. RFID can provide benefits such as operational efficiency, improved visibility, reduced cost, improved security, improved customer service levels, better information accuracy and increased sales. However, the improved product visibility and operation efficiency are the most significant benefits (Bhattacharya, Chu, & Mullen, 2007). Product visibility provides the ability to eliminate inventory discrepancy, meaning that product visibility affects the reduction of unnecessary disposal of food (Angeles, 2005). Furthermore, RFID systems contribute to the optimization of operational processes of supermarkets by speeding up and removal of certain operations.
3.3.1.1 Product visibility
Supermarkets have to deal with several activities that can result in inventory discrepancy, which leads to inefficient inventory management, resulting in higher inventory levels or poor customer service (Atalı, Lee, & Özer, 2006). Currently, without RFID, supermarkets’ inventory policies and control systems are managed based on recorded inventory, while the physical inventory may differ from what supermarkets expect (Lee & Özer, 2007). Ultimately, inaccurate inventory records leading to inaccurate replenishment orders, resulting in excess inventory that expire over time. However, RFID technology is capable of solving this information accuracy problem, while reducing inventory levels and lowering distribution and handling costs by providing accurate and real-time information of inventory quantities and movements (Jones, Clarke-Hill, Hillier, & Comfort, 2005).