Reflad

[IMPORTANT] Note: The content on this page is an English translation of content based on a proposal submitted to a business-idea competition in Japan in August 2022. The content on this page therefore does not necessarily capture the state of the project or the ideas behind it at the time of the creation of the translation in May 2024.

1. Title

A retail-management system feature to promote temaedori* by consumers, 

thereby reducing food waste, reducing inventory loss, and increasing revenue for retailers: 

a collaborative endeavor

3. Background, motivation, and thesis

   Climate action is one of the UN’s sustainable development goals. The Paris Agreement, adopted in 2015, emphasizes the importance of keeping global air-temperature rise less than 2 degrees centigrade compared to pre-industrial levels. To this end, 120 countries around the world share an awareness that we need to achieve net-zero emissions by 2050. While progress is being made in many areas toward this goal, the road ahead is by no means easy. 

   In Japan, certain companies and other organizations with forward-looking perspectives and advanced techniques have made significant progress. However, some say that there is still a long way to go when we look at society as a whole. There are some experts who say that there are many technologies that already are available to achieve net zero. The progress we make from here will depend on how effectively we can utilize those technologies to reduce carbon dioxide emissions.

   If we could convert all sources of energy required for human activity from fossil fuels to renewables, then we can essentially achieve zero CO2 emissions. However, there is still a lot of energy used worldwide that depends on fossil fuels. Therefore, until the complete conversion to renewables is achieved, efforts to reduce energy consumption across all areas of our lives will result in the reduction of the overall amount of fossil fuels we consume. One such area of our lives is the food system. Unsurprisingly, the food system requires energy in all stages of the supply chain, from agriculture to processing, logistics and retail.

   If we suppose that there are ten units of food that are produced worldwide, and three of those units are being lost or wasted, then approximately 30% of the energy that is being used to produce and supply the ten units of food is also being wasted. Therefore, in the ultimate case where we succeed in achieving zero food waste, we can reduce the amount of energy used throughout the food supply chain by about 30%. Food waste worldwide is estimated to contribute to about 6 to 8% of all CO2 emissions. Therefore, if we can reduce food waste, then we can contribute to reducing carbon dioxide emissions and therefore mitigating climate change. If we manage to achieve zero food waste, we can reduce global carbon dioxide emissions by 6 to 8%. While achieving zero food waste might not be too realistic, a scenario in which we reduce food waste by 50% will allow us to reduce global carbon dioxide emissions by 3 to 4%.

   So now, let's talk about some topics in addition to climate change. In many developed economies, including Japan, aging populations and declining birth rates are a major social problem. In Japan in particular, we hear about the importance of promoting digital transformation and boosting productivity as labor becomes increasingly scarce. As burdens increase on the healthcare system, efforts are being made to increase healthy life expectancy and promote preventive medicine. This includes taking care of our physical and mental health and well-being in our daily lives. In relation to this, we see work-style reforms being advanced throughout society. In this context, the product idea we are proposing targets not only climate action and food-waste reduction but also aims to contribute to work-style reforms.

   When we look at what's happening in society, even when we say that climate action is important, society as a whole tends to direct its attention to dealing with problems whose consequences materialize in the shorter term. These relate to issues such as infectious diseases including COVID-19, economic policies, and national security. Businesses, too, tend to show little attention to products or services that solely emphasize climate action. Therefore, we think it is important that we propose a package of benefits that provides benefits in addition to climate-change mitigation. The proposed product idea contributes on a societal level by reducing food loss and mitigating climate change. On the other hand, for businesses, the product idea reduces lost inventory arising from expiration of food, and achieves improved employee well-being and job satisfaction through increased productivity. Furthermore, the product idea benefits consumers in both financial and non-financial ways. Financially, consumers spend less on their groceries. In addition, the product idea is expected to bring about increased self-esteem and self-image in the consumer, who engages in ethical purchasing behaviors that involves opting for close-to-expiration food items. This product idea thus provides benefits to multiple stakeholders.

   In terms of work-style reform, one way to achieve improved job satisfaction is to free the employee from a sense that they are doing work that is entirely repetitive and let them develop a sense of pride within that they are doing what only they can do and cannot be done by machines. As an end goal, an ideal workplace culture would involve the employee feeling pride or joy about the work they are doing at every moment. In such an environment, in addition to making efforts to smile, the employee might find themself smiling or laughing as an effortless, natural response to an interaction they might have with a coworker. There might also be a spontaneous conversation or joke followed by laughter with a customer, too, when circumstances permit.

4a. Overview of the product idea—with emphases on uniqueness and competitive advantage

     Our product idea is to contribute to food-waste reduction through an advanced inventory-management and payment system for food retailers such as grocery stores and convenience stores selling delis/prepared foods and other perishable products that have relatively short shelf lives ("DPFOPPs"). This system will be configured as a feature within the larger retail-management system that a typical food retailer already has in place. Our product idea involves performing two main steps before the products are put on the store shelves: (1) entering data on expiration date (and time) (“EDT”) into the inventory-management system and (2) implementing a set of algorithms setting forth a predefined discount rate corresponding to time-to-EDT at checkout for each product type. Once these steps have been performed, a close-to-expiration food item will no longer require a staff member to locate the item and place a discount sticker on it to prevent it from remaining unsold past its expiration. Instead, a shopper can simply bring the item to the cashier, scan the barcode, and pay a discounted price.

   Initiatives based on similar ideas are not entirely new. In Japan, (a) 7-Eleven has rolled out a project called “Ethical Project,” and (b) the supermarket chain Ito-Yokado performed a pilot project in January to February of this year [1]. (c) Furthermore, by 2025, 7-Eleven and other major convenience store chains plan to introduce RFID tags [2]. The above (especially (a) and (b)) are cases where one or two leading vendors in the industry are involved in a forward-looking initiative, demonstrating how the situation explained in the second paragraph of Section 3 is occurring in the food industry. We want to achieve net zero as a nationwide effort, and ultimately as a planetary effort, by 2050. Bearing this in mind, even though an initiative with a similar concept has been pioneered and implemented elsewhere in the food industry (by Ito-Yokado and others), engaging widely with businesses across the food industry to promote the implementation of similar initiatives and building a track record of efforts toward net zero seems to be a meaningful endeavor.

   Large vendors may have control over their entire supply chains. On the other hand, it may not be easy for small and medium sized retailers to manage their products from earlier on in the supply chain. Bearing this in mind, it is important that multiple actors work together across industries and sectors (industry, academia, government, and civil society [3]). To differentiate our initiative from the initiative of Ito-Yokado and others, which is an example of a private–public partnership, we think it is important that we emphasize collaborations with academia and civil society.

   In addition, we intend to roll out an initiative that not only emulates and disseminates what Ito-Yokado and others did through their pilot project to other parts of society but also features additional elements that were not included in the pilot project (see Section 4b on details of the product idea). 

   We intend to include an element of public good in the product of our endeavor. Because we are aiming to promote multi-actor collaborative innovation, the product of this endeavor will be a result of contributions from multiple stakeholders regardless of the size of the contribution or whether the contribution is direct or indirect. It would be ideal if the product of this endeavor benefited those who were not necessarily directly involved in the development of the product. For example, if the product of the endeavor could be made available for download from an open-source repository like GitHub and could be added as an add-on feature to the POS system that a vendor is already using, then there will be the prospect of widespread adoption and use. More details on monetization will be provided in a later section.

4b. Details of the product idea 

   This project will be divided into three phases: Phase 1, Phase 2, and Phase 3. The classification of DPFOPPs into three product groups is an important part of describing the three phases.

• [Group 1] These products have shelf lives of 1–2 days and expiration data for both date and time (e.g., bentos, raw fish, some delis) [relevant to Phase 3]

• [Group 2] These products have shelf lives of about 3 days to 1 week and barcodes that are printed in-store (or remote in-house locations) (e.g., seafood/meat products prepared in-store, some delis) [relevant to Phase 1]

• [Group 3] These products have shelf lives of about 3 days to 3 weeks and barcodes that are not printed in-store (e.g., bread, milk, juice, some delis) [relevant to Phase 2]

   Here we will focus on Group 2, which is the primary focus of Phase 1. For this product group, the discount rate will not change in the middle of a given day. This way, a shopper will be unable to put a product in their cart and wait for the discount rate to change before bringing the product to the checkout counter. Another factor that makes this product group relatively easy to work with is that the expiration data linked to the barcode for each item can be managed in-store. 

   The family of products making up Group 2 is not necessarily exceptionally large within a grocery store’s entire product line. However, owing to the relatively short shelf lives, these products need to be monitored by employees at relatively short intervals. By automating some of the markdowns for this product group, more employee hours can be committed to Group 1 products, which have even shorter shelf lives and therefore require price management on the time scale of hours. Such an arrangement will enable grocery stores to maintain their systems of food safety while minimizing inventory loss due to food waste.

[Concerns]

   A potential concern relates to the issue of product display. How will a shopper be able to tell apart a Group 2 product (which is processed in-store) from a Group 3 product (which is processed elsewhere) if they are on the same shelf? It is quite possible that both of these products will be on the same shelf because they both have shelf lives of about 3 days to 1–2 weeks. 

   This concern will be addressed first by using product labels such as those shown in Figures 1–4 for Group 2 products. These labels will each have a unique color and pattern [4] (collectively, “background design”). Four background designs have been defined to allow for four subdivisions of products within Group 2. For each background design, there is a corresponding banner for the label. The banner shows a chart showing the relationship between the remaining shelf life and the discount rate for this Group 2 product subdivision. The banner will be placed in the middle of a shelf or on the front door of a refrigerated display case containing a Group 2 product with a label with the same background design (Figure 5).

   Let's explain next why we have four different background designs. To do this, we will use some specific numbers. Suppose there are two products: Product A has a shelf life of three days, and Product B has a shelf life of seven days. The discount rate to apply to the two products may differ at two days before their respective expirations. That is, Product A has barely been on the shelf (perhaps only several hours have gone by since it was placed on the shelf in the morning), at two days before expiration—perhaps in the late afternoon. Therefore, the discount rate will likely be zero. On the other hand, Product B has been on the shelf for four days at two days before its expiration. Therefore, it might be reasonable to apply a discount of a certain amount. To account for this type of difference that might occur between products with different shelf lives, we thought it would be helpful for the shopper if they could see the discount rate to expect for each product with a different shelf life (“subdivision”) at different stages of its life on the shelf. We suppose a case in which a chart and a corresponding label with the same background design is defined for each value of n, where n is the shelf life of a Group 2 product and takes a whole number between 3 and 7. This means that 4 or 5 background designs are required. Figures 1–4 show the label and the corresponding chart for each background design. 

   Let's now look at this system from the shopper’s point of view. The benefit of this system to the shopper in a nutshell is this: When I, a shopper, see a product with one of the background designs at a grocery store, I can just purchase that item if I know at that point that I'm going to be eating that item soon. By engaging in that kind of purchasing behavior, not only will I be able to save on my spending but I will also be able to boost my self-esteem that I am engaging in ethical purchasing behavior.

   Let's look further into how this system might work from the shopper’s point of view. I walk along an aisle in a grocery store and stop in front of a shelf like that shown in Figure 5 with a collection of charts with different background designs. On seeing that, I will understand that products within that shelf with labels with the same background designs are eligible for discounts [provided that I am aware of this novel system through an advertisement I have read elsewhere]. To get an idea of the rough discount rate I can expect on a product I pick up from that shelf, I will first subtract the expiration date from the current date and calculate the remaining shelf life for that product. I will then find the row in the chart with the same background design as the label on the product that shows the remaining shelf life in the left column. The range of percentages in the right column of the same row tells me the rough discount rate for my product. But I want to know exactly how much I would be paying. For customers like me, it would be useful to have a scanner and a tablet device near the shelf. That way, I would be able to scan the barcode for my product right in front of the shelf and ascertain the exact price I would be paying. 

   The system we propose also allows for adjusting the discount rate by remaining shelf life (where [remaining shelf life] = [shelf life] – [days already spent on the shelf]), for products of the same subdivision. For example, if the shelf life is 3 days, it is possible that some products will undergo very little reduction in quality after one day on the shelf. A different product might undergo significant reduction in quality during the first day on the shelf. Given such contrasting products, it would be reasonable to assign different discount rates to be applied at two days to expiration. To enable such accommodations, each row on each chart displays a range of discount rates: e.g., 1–9% (see Figures 1 to 4).  The charts here are assumed to be printed on paper. However, it may be better to show the charts on flat-panel displays to enable frequent modifications in the relationship between remaining shelf life and discount rate. 

   For Group 3 products, Reflad will stick with current practices in which single-color discount stickers are placed manually on eligible items, as shown in Figure 6. This will allow customers to tell Group 3 products apart from Group 2 products visually. 

[Partial automation/hybrid implementation] 

   It is also worth noting that Reflad does not require the markdown process for Group 2 products to be fully automated. If employees are accustomed to making the rounds and checking food items manually one by one to ensure that they do not remain on the store shelves past their expiration dates, such a practice can be maintained in conjunction with the implementation of Reflad. What Reflad will enable is that it will automate the process of applying small discounts of less than 10% at frequent intervals from earlier on in the life of a product after it is placed on the shelf. We hypothesize that this mechanism will reduce the probability that the product will remain on the shelf until there is only one or two days left to expiration, at which point the decision to apply a 10 or 20% discount will become inevitable. 

   If Reflad does indeed succeed in achieving the aforementioned kind of reduced probability, there should be a reduction in the frequency at which employees have to print and place discount stickers on food items when they make the rounds. For products that only have one or two days left until expiration, a grocery store can choose to apply dedicated discount stickers on top of the labels with the background designs, if the presence of the dedicated discount stickers is expected to more effectively attract customers’ attention. The benefit of implementing Reflad in this hybrid manner is that there will potentially be less confusion arising from changes from current practice. In this case, the markdown automation will only be applied in cases where there are still a few days left until the expiration and the discount rate to be applied is less than 10%. Figures 1 to 4 show the charts when Reflad is implemented in this hybrid fashion. The box at the end of the main text provides an example with hypothetical numbers. 

5. Who will this target?

   This proposal describes an idea for a product to sell to businesses, namely businesses in the food industry. Moreover, the food-industry businesses we are targeting are not B2B-oriented (such as wholesalers) but B2C-oriented: i.e., retailers targeting consumers, such as supermarkets and convenience stores. The reason for this is that B2B-oriented businesses in the food industry have a fairly good grasp of their customers’ demand. Therefore, they have often already optimized the timing and volumes of their purchases. This kind of reasoning can also be applied when comparing different types of B2C-oriented businesses. Let’s take the example of a food-service provider that delivers bento meals to retirement homes. The number of residents in a retirement home to which the bento business caters is not subject to frequent fluctuation. Based on this relatively stable number of residents, the bento business knows that it will receive orders of fairly consistent volumes at fairly consistent intervals. Such a business already has a good understanding of the volumes that it will supply in such a way that it matches demand. Therefore, this kind of business does not need to worry so much about food loss and waste.

   We are thus narrowing down our focus to B2C-oriented businesses such as supermarkets. But still, the needs of each business are different and depend on the size, content, and other characteristics of the business. For example, let’s think of a business that is reasonably small and whose customer base largely consists of loyal customers who shop on a fairly regular basis. (Relatively upmarket vendors such as SEIJO ISHII and KINOKUNIYA come to mind.) It is also possible that the business deals in a relatively specialized line of products. In such a case,  it is possible that the management already has a fairly good understanding of customer demand, and the probability of products remaining unsold past their expirations is not a large concern. And even if such close-to-expiration products do occur, their volumes are not so large. Therefore, the burden for employees to place discount stickers on the products individually is not so heavy. For this kind of business, there is a reasonably good match between its supply and customer demand. The need for a food-waste solution is not so strong. 

   Based on the above, we envision that some of the early adopters of our product will be food retailers that sell large volumes of products, both in terms of product variety and volume per product, and have broad customer bases. Accurately predicting customer demand is not always easy for these businesses. There exists the possibility that a fair amount of products will remain unsold past their expirations. A significant number of employee hours will potentially have to be invested to prevent this possibility from materializing. 

   The kinds of supermarket that meet these conditions that first come to mind are large- or medium-sized chains that cater to broad customer bases centered on middle-income consumers. Prominent vendors in the northwest of Chiba Prefecture, where we are based, are Tokyu Store, AEON, and York Mart. We would first like to ask such vendors for their cooperation in testing our idea. 

6. Timeline for operationalization; monetization

[6.1. Phase 1]

   Efforts relating to Phase 1 will be our main focus for the period leading up to the introduction of RFID tags: i.e., from 2022 to 2023 or 2024. The skills required for this phase will center on programming. Sophisticated ICT or biotech will not be required. We will aim to implement an augmented version of what was tested by Ito-Yokado and others in their pilot project, in a wide range of supermarkets and convenience stores, including small- and medium-sized vendors as well as large vendors. Specifically, we would like to start discussions with a potential partner to conduct a pilot project with by the end of 2022. After completing a successful pilot project, we will aim to enter into our first paid agreement by the end of 2023.

[6.1.1. Monetization]

   We intend to employ a performance-based revenue model, in which we receive a portion of the reduction in expenses that an implementer of our system succeeds in achieving [5]. Especially during the early stages of operationalization, we will have to address the problem of how to cover the upfront costs of installing the system. These costs involve hardware such as monitors and electronic shelf labels. We are considering the possibility of launching a crowdfunding project that specifies the use of funds received. 

   After entering into the first paid agreement in 2023, we will start advancing efforts for increased capabilities and features, such as developing mobile applications as needed, at around the same time as when efforts relating to Phase 2 will commence (see the next section). 

[6.2. Phase 2]

   Efforts relating to Phase 2 will commence in 2023 and will last for about two years. 

   When Group 3 products—the focus of Phase 2—are received at a supermarket, they already have barcodes printed on them. The barcode format currently used by many food retailers is the JAN format. This allows for 13 digits. Other barcode formats exist, which allow for the storage of more information. One such format is GS1-128. If such a barcode standard is adopted across the supply chain, it will be possible to create an integrated system that associates a unique piece of data for the expiration (date (and time)) of each product lot at the time of its manufacture. The same data will stay associated with the product lot throughout its entire journey through the supply chain. 

   A key challenge in Phase 2, then, will involve establishing a multi-lateral agreement to print product-lot-specific barcodes. The printing will be performed by actors further up in the supply chain, who will be supplying the products to downstream actors, such as supermarkets.

   Adopting a new barcode format across the entire sector, however, will require a sector-wide movement, just like several decades ago when the current barcode system was adopted. To achieve this, strategic efforts will need to be made over a significant amount of time. A good time to start planning for this might be after the first paid agreement associated with Phase 1 gets underway. (This will also coincide with when Phase 3 starts.) At that time, we would need to do research on how the utilization of standards such as GS1-128 has fared in regions where they have been adopted, such as Europe. 

[6.3. Phase 3]

   In Phase 3, we intend to utilize RFID tags. As such, Phase 3 will likely require more spending relating to the procurement of RFID tags and other equipment, than in Phase 1. More rigorous analyses of cost versus benefit will be required. Here again, collaboration with academia will play an increasingly important role. An effective approach might be to develop and utilize tools such as an agent-based model and do a cost-versus-benefit analysis in which we simulate the amount of time and trajectories that a shopper takes to bring a discount-eligible item from a store shelf to the checkout counter. We are also interested in considering the possibility of playing a role (by, for example, participating in the bidding process) in the government-led effort to promote the widespread use of RFID tags.

7. Vision for the future—looking beyond the scope of this proposal

   This proposal deals only with Phases 1–3 of our project. In “Phase 4” and beyond, we would like to consider increased capabilities and features that utilize biological and/or chemical techniques, in addition to the information and communication technologies that form the core of efforts up to Phase 3. For example, a technology exists to prevent counterfeit pharmaceuticals that involves adding microparticles to the surface of the medicine [6].  If we can apply this technology to develop a technique to determine the amount of time that has passed since fruits and vegetables have been harvested, it might be possible to adjust prices according to the freshness of the produce. 

Notes

[1] https://www.meti.go.jp/press/2021/01/20220111004/20220111004.html

[2] https://warp.da.ndl.go.jp/info:ndljp/pid/11253807/www.meti.go.jp/press/2017/04/20170418005/20170418005.html

[3] Here we envision “civil society” as the sector of society consisting of NPOs and other organizations that more prominently feature the voices of individual citizens.

[4] To accommodate color-blind users

[5] Once we achieve a certain level of revenue through agreements with medium-sized/large enterprises, we should be able to make the system available for download and use free-of-charge as an open-source module described in the section on the overview of the product idea. This should broaden the range of users. Specifically, increased capabilities should emerge to automate the inventory management of smaller specialty-oriented businesses that deal only in certain types of products, if requested by the management.

[6] https://trutags.com/

[Box] 

   Let's look at a specific example. Suppose there is a product (Product A, such as an okonomiyaki) that has a shelf life of 5 days ([date of expiration] – [date of production]).  We assume that the grocery store selling this product currently places a 10% discount sticker when an item on the shelf selling this product remains unsold at 2 days before expiration and applies a 20% discount sticker on the day before expiration. Our proposed system will enable the option of automating the application of 1% discounts at 4 days before expiration and 2% discounts at 3 days before expiration.

   Now let's imagine a simple situation in which there are only two items on this shelf: one with 5 days to expiration and another with 3 days to expiration. Currently, before system installation, both of these items are being sold at the full price with no discount. Suppose there is a shopper who currently preferentially picks up the item with 5 days to expiration. After system installation, however, this shopper learns that they can purchase the item with 3 days to expiration at a 2% discount. On learning this, when this shopper has no real need to buy the item with 5 days to expiration, there's a good chance* that they will buy the item with 3 days to expiration.

[* This “good chance” could be quantified on the basis of scientific evidence by collaborating with academia and developing, for example, an agent-based model built around behavioral-economic principles.]

   In this situation, currently the item with 3 days to expiration remains unsold until the following day. After system installation, however, this item is purchased instead of the item with 5 days to expiration. This means that there will be one fewer item on the shelf that remains unsold until 2 days to expiration, compared to current conditions. This means that, when an employee makes the rounds to apply discount stickers on any items that have two days or less to expiration, there will be one fewer item requiring a discount sticker.

   Below we present a rough idea of the economic benefit of installing our system by comparing the flow of money before and after system installation, focusing on the period of a few days after the item described above is placed on the shelf until it is sold. 

   The revenue is the full price of this one item both before and after the installation of the system. The costs before system installation are the cost of labor for the employee to print and place the discount sticker and the cost of the 10% discount (assuming this item is purchased 2 days before the expiration).  Meanwhile, the costs after system installation are the cost of the 2% discount. 

   Let's plug in some numbers. Let's suppose that the employee who prints and places discount stickers while making the rounds earns 1500 yen per hour. If this employee takes 15 seconds to print a barcode sticker and place it on the relevant item, the cost of labor for this one operation is about 6 yen. If the full price of this item is 400 yen, a 10% discount is worth 40 yen. Therefore, the costs before system installation amount to 46 yen. Meanwhile, the costs after system installation are equal to 2% of 400 yen. That's 8 yen. The costs after installation are much less than before. This demonstrates the economic benefit of the system.

   Looking ahead, however, more rigorous calculations will be required. There is the need to include, as part of labor costs, the cost of associating expiration data for this and many other items with their respective barcodes and entering these into the database. In addition, estimating the cost of labor for the employee will likely be less straightforward than calculated above. Therefore, a comparison of observed values (collected through a pilot project) and theoretical values obtained from agent-based modeling founded on behavioral-economic principles through collaboration with academia would be an effective approach toward operationalization. 

Figures