Implementation of RFID in food supply chains

In this section, we intend to predict the number of firms which possibly adopt the RFID technology in food supply chains. Because the food traceability system is a new technology, it takes a relatively long time for firms to make the adoption decision. In order to predict the number of firms implementing a food traceability system, we focus on the theory – diffusion of innovation (Kolter, 1994). Innovation diffusion theory has been widely used to understand technology adoption in organizations (Ranganathan and Jha, 2005; Lee and Shim, 2007; Sharma, Thomas and Konsynski, 2008). The theory provides a useful perspective on how to improve technology assessment, adoption and implementation (Matta and Moberg, 2006). Rogers (1995) indicates that when the numbers of consumers who adopt a new technology are plotted over time on a frequency basis, they follow a normal and bell-shaped curve. In addition, when the cumulative numbers of adopters is plotted over a length of time, they follow an S-shaped curve. The diffusion of innovations is shown in Figure 2. The bell-shaped curve shown in Figure 2 is simply divided into five sections with the different standard deviations of the normal distribution. As the S-shaped curve shown in Figure 2, with consumers adopting the new technology, its market share eventually reaches the saturation level.

Figure 2: The diffusion of innovations (Rogers, 1995)

Rogers (1995) suggests that one may utilize an S-shaped curve to illustrate the diffusion of innovation. The S-shaped curve rises slowly when there are few adopters at initiation. It then accelerates to a maximum until half of the individuals in the system have adopted the new technology. Then the S-shaped curve increases at a gradually slower rate as fewer and fewer remaining individuals adopt the innovation. In practice, the logistic distribution gives a nice looking S-shaped curve with a relatively simple mathematical formula (Stockute et al., 2006). It follows that our study utilizes the logistic distribution to approximate the S-shaped curve.

Balakrishnan (1992) indicates that the logistic distribution is applied to various growth models especially in describing the population and organismic growth. The cumulative function of logistic distribution has been used as a growth curve. As a result, we utilize the logistic distribution to estimate the growing number of firms in this paper. Let F(x) denote the cumulative distribution function (cdf) of the logistic distribution with variable x, which is stated as follows (see Stockute et al., 2006):

(1)

In addition, the probability density function (pdf) of the standard logistic distribution is given

The conventional food supply chain includes food manufacturers, distributors and retailers. As soon as the goods are produced, they should be labeled an RFID tag and the production information should be recorded, so the plants of each food manufacturer and the distribution centers have to install the RFID equipment in the beginning. On the other hand, there is a large number of retailers in the market. It would take a certain time to promote the RFID technology to the entire food supply chain. We assume that the retailers would adopt RFID following an S-shaped curve in T time periods. We let N denote the total numbers of food manufacturer plants, M denote the total number of distribution centers, and L denote the total number of retailer stores. We consider that there are lt retailer stores adopting the RFID technology in the t^th time period. To estimate lt and consider the fluctuation, we use the

According to the above estimation procedure, we can derive the number of adopters including

food manufacturers, distributors and retailers in food supply chains within T time periods.

4.2 The pricing strategy

In order to commercialize the food distribution traceability system in food supply chains, we work out a plan of pricing strategy. The adopters of food distribution system are thought as clients who install RFID equipment. In practice, there are three common approaches to pricing: cost-based pricing, customer-driven pricing and competition-driven pricing approaches (Collins and Parsa, 2006). Cost-based pricing simply means that the price is based on the costs of the product and an equitable profit. Customer-driven pricing is a pricing approach that is determined by market demand and the different reaction of consumers.

Competition-driven pricing is a market driven pricing approach that is determined by the market price of similar products of the competitors and production conditions. Among these three pricing approaches, cost-based pricing is relatively easy to implement because it requires little information and is easy to calculate. In addition, the cost-based pricing can be further divided into several pricing approaches. One of the approaches is breakeven pricing, which is usually used when the number of sales is established. In breakeven pricing approach, prices meet some certain standards in order to achieve the breakeven point. In this pilot project, we adopt the breakeven analysis to pricing because it could determine a breakeven point for firms to a balance between receipts and payments.

Breakeven analysis is a technique used to identify the sales volume at a price that covers costs. The breakeven point is the level of sales at which total revenues equal total costs. In order to find the breakeven point, we estimate the total costs and the revenues that would be earned at various quantities at a given price. Both the total cost and revenue are linear functions based on the simplifying assumptions that (i) each unit is sold at the same price, and (ii) the costs of producing each additional unit is the same.

ven oint otal Fi ed osts

rice nit ariable ost (6)

In (6), the Total Fixed Costs are the costs that do not vary with the level of production quantity. In other words, for example, the total fixed costs are the same whether a firm produces one unit of their product or one million units. Fixed costs typically include the expenditures such as the rent on the building in which the firm produces its products. On the other hand, Unit Variable Cost is the cost varying with the level of production quantity.

When we estimate the breakeven point, the sales volume needs to cover costs at a specific price. In addition, it is necessary to consider whether the firm is likely to sell the quantity at the given price. Therefore, we combine breakeven analysis with knowledge of consumer demand. With considering a desired profit into the estimated breakeven point, we can incorporate a specific profit objective into breakeven analysis to generate a desirable profit.

Thus we can determine the sales volume to achieve the desired profit level.

ales olume

(7) Consequently, from (7), we can set price as (8) which simultaneously considers cost, demand factors and produce valuable input. start-up cost includes fixed assets such as fixtures, equipment, decoration, office supplies and installation of equipment. The operations costs include rent, payrolls, insurances and utilities. We let TCt denote the total cost of time period t. Similarly, the TSC denotes the total start-up cost and TOCt denotes the total operations cost of time period t. Thus, the total cost at the first period is the sum of the total start-up cost and total operations costs of the first

time period. In the remaining periods, the start-up cost does not take place from the second period. More specifically, the total cost can be stated as

(9)

Based on (9), we can calculate the resulting costs of the ASP for promoting the RFID to each member in a food supply chain. Both the start-up cost and operations costs are classified as fixed costs because they do not vary with the different number of customers. At the same time, this ASP provides customers with the service of hardware installation and consulting service. Compared to a manufacturing industry, a service industry makes profits by offering service to customers. It is difficult to formulate the variable cost with a new customer, so a service industry usually does not mention the variable cost.

Based on breakeven pricing of (8), we can set a price on the RFID service. The RFID hardware is provided by the RFID developer and installed by the ASP. In this research, we focus on the cost and profit analysis of the ASP. The procurement cost of the RFID hardware is eventually transferred to clients installed the traceability system. In this pilot study, we charge the customers an initial fee, named Pinitial, paying for installing the RFID-related hardware or equipment. On the other hand, charging Pinitial is to make the breakeven point of start-up cost of the ASP. As a result, Pinitial can be derived from dividing the start-up cost by the total number of clients, which consists of food manufacturers, distribution centers and retailer stores. That is, the Pinitial can be generally stated as

(10)

In addition to the initial fee, we may charge customers a service fee after the first installing period. We let Pservice denote the service fee. For the purpose of making profits, we multiply operations costs by a proportion of desired profit x%, and then divide it by the number of accumulated clients for service period T. For instance, if x% is equal to zero, the

business would reach breakeven. As above, the service fee can be formulated as following periods. In the following, with the information of price and number of clients, we can calculate the profit of this pilot project. The purpose of calculating profits is to understand the potential revenue of a food traceability system.

4.3 The expected profitability

In the preceding sections 4.1 and 4.2, we illustrate the method of setting price and forecasting number of customers. Now we can estimate the profit of this pilot project. We simply set that the total revenue is price multiplied by quantity. The profit is calculated by revenue minus costs, which is stated as shown below

. We note that the RFID equipment of hardware is offered by the RFID developer and supplier.

The ASP simply provides the service of installing the system and information services.

Because the ASP simply transfers the procurement cost of the RFID equipment to its clients, the procurement cost is not included in the analysis in this study. According to our study, the food manufacturers and distribution centers adopt the RFID technology in the first time period. On the other hand, the adopting number of retailer stores would increase from l1 to lT

in T periods. As a result, profit of each period can be stated as follows:

adopting the RFID in the preceding period are charged a service fee. Because the service fee is charged from the second period for a single client, the price would make the breakeven point in period T+1.

In this section, we develop a method to forecast the clients and set the pricing strategy of this pilot project. In the next section, we apply these developed methods to a field study.

In our opinion, the chain convenience stores in Taiwan may be a suitable choice to run the traceability system in the beginning because of their standardized business methods. In addition, these convenience stores usually have their own distribution centers to support the goods delivering. As mentioned earlier, to improve the viability in commercializing the food distribution traceability system, we conduct a field study for chain convenience stores in Taiwan.

Chapter 5. The Field Study

5.1 Data collection from the chain convenience stores

The distribution traceability system serves a role to help us enhance the food safety in a food supply chain. To implement the task, the pilot project runs the food distribution traceability system to link itself with a set of suppliers, distribution centers (DCs) and retailers including convenience stores, supermarkets and warehouse stores. For the reason of viability, we restrict the study range to chain convenience stores in Taiwan. They have adopted the electronic commerce system so that it is easier for them to implement the food/product traceability system by using RFID technologies. As expected, launching the RFID technology results in customers with convenience and security. In addition, we note that the major groups of chain stores in Taiwan are 7-Eleven, FamilyMart, Hi-Life and OK-mart.

The target markets for the RFID are convenience stores whose main customers expect fresh, convenient, nutritious and inexpensive products.

The goal of this pilot project is to promote RFID systems to chain convenience stores, including the retailer stores, distribution centers and food manufacturers. Once the goal of implementing RFID system to convenience stores is successfully reached, it can be extended to the next stage of supermarkets and warehouses to apply the RFID system.

For analyzing purposes, we have to measure the market size and the demand of RFID hardware/software in two aspects: one is to estimate the number of equipment needed in food supply chains, the other is to verify the target market to promote the RFID technology. We first calculate the number of required instruments of RFID readers and tags by the following estimated approach. Note that there are different types of RFID readers including the fixed reader, the hand-held reader, the RFID-POS system and the kiosk. In a similar manner, the demand of readers can be discussed in two sides: manufacturers and retailers. For a

manufacturer, fixed readers are needed in order to record the information of products. For retailers, each retailer store has an RFID-POS at checkout counter to record the selling information and a kiosk for consumers to inquire the relative data of their purchases. The retailers and distribution centers also need hand-held RFID readers for stocktaking. For a systematic analysis, an integrated procedure is proposed and contains three stages as shown in Figure 3. The first stage is to investigate the chain of convenience stores in Taiwan. In the second stage, we sample and inquire the convenience stores for estimating the number of purchases. On the other hand, we investigate the number of convenience stores, major food suppliers and their related distribution centers. Applying the analyzing process, we can estimate the consumption of RFID tags and the requirement of RFID equipment in the convenience store chain.

Figure 3: The procedure of estimating the number of RFID equipment

It is worth to mention that each chain of convenience store has its own logistics system.

Obviously, the needed numbers of readers, RFID-POS and kiosks can be estimated by the firm of convenience stores. With the information gathered from each convenience store, the number of stores and related distribution centers in each convenience store chain is shown in Table 4. In addition, some larger food manufacturers and the distribution centers of each firm are summarized in Table 5 as well.

Table 4: The number of stores and related DCs of each convenience store

Table 5: The number of plants and related DCs of major food manufacturers Firm Number of plants Number of DCs record the information of inventories of merchandise, two RFID-POS systems at counter for clerks to checkout and a kiosk for customers to inquire the information about goods. In addition, each food manufacturer has at least two fixed readers to record the production and transportation process. Moreover, two fixed and one hand-held reader assumed to equip in a distribution center. As above-mentioned statement, the requirement of RFID equipment is shown in Table 6.

Table 6: The requirement of RFID equipment to each member

Under this consideration, the amounts of fixed readers, hand-held readers, RFID-POS systems and kiosks can be calculated and are summarized in Table 7.

Table 7: Estimated equipment demand of distribution traceability system Fixed reader Hand-held reader RFID-POS system Kiosk

Food manufacturers 60 30 N/A N/A

In the following, we calculate the consumption of RFID tags in chain convenience stores.

We can estimate the consumption of RFID tags by the field data from retailer stores. We take samples from five different locations of convenience stores in Taiwan. The purchase data are recorded in two time periods, morning and evening, which are from 10:30 to 11:30 in the morning and from 7:00 to 8:00 in the evening. There are four kinds of foods that are suitable for utilizing RFID tags. Those are beverage, instant food, packaged food and frozen food due to their characteristics. We record the number and types of purchases by observing customers of the checkout counter. In addition, we summarize the resulting data in Table 8.

In Table 8, each cell indicates the consumption of each kind of food during a specific time period.

Table 8: Records of tags consumption for firms

Package or Morning Evening Morning Evening Morning Evening Morning Evening Morning Evening

Plastic bottles 11 29 48 48 25 12 17 63 20 16

From the available data, we summarize the whole number of items in each specific store given the time period. Some outlier statistics are deleted as shown in the cross-out marks in Table 8 (i.e. 132, 104, 76 and 42); that is, we delete the highest and the lowest number of morning and evening respective. With the field study data, we can estimate the day consumption of tags in each retailer store by

hrs tags day Thus, the total annual number of tags used can be approximated as

days stores

With the former estimations of required RFID readers and predicted amounts of tags, the potential market of hardware is large according to our analysis. Because the number of tags consumption is 6,426,079,040 per year, it is worthy of investing in this field.

5.2 Promotion scenarios of food traceability system

By the illustration of section 4.1, the adopting process of the RFID technology in chain convenience stores would follow a specific sequence. In this field study, we take a year as a planning time period. The food manufacturer and distribution centers should install the RFID equipment in the first year. The retailer stores may install the hand-held readers and kiosks in sequence. Following industry experts’ opinions, we investigate two scenarios in the study: 4- or 6-year promotion schemes. The adopting numbers of retailer stores follow the standard logistic distribution as shown in (5) in section 4.1. We let M denote the number of food manufacturers, N denote the number of distribution centers, and L denote the number of retailers, where M is 30, N is 45, and L is 9,200 in the field study respectively. This helps us derive the result shown in Figure 4 and Figure 5. For 4-year promotion scheme, the accumulated clients of retailer stores installed the RFID-based traceability system through the first year to fourth year are 2,000, 4,600, 7,200 and 9,200. On the other hand, for 6-year promotion scheme, the accumulated clients of retailer stores installed the RFID-based traceability system through the first year to sixth year are 1,300, 2,600, 4,600, 6,600, 7,900 and 9,200. This allows us to approximate the suggested price in a reasonable manner. As expected, once the pilot project is successfully promoted throughout the entire convenience store chain, the RFID traceability system then may be expanded to other fields.

Retailer stores 2,000 4,600 7,200 9,200

0

5.3 The cost and profit analysis

To set the price of implantation of RFID, we estimate the annual expenditure including start-up cost and operations costs. Because salary cost is one of the most important parts of

To set the price of implantation of RFID, we estimate the annual expenditure including start-up cost and operations costs. Because salary cost is one of the most important parts of