SID5_ft0348 - Secretary of State for Environment, Food...

General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: [email protected]

SID 5 Research Project Final Report

SID 5 (Rev. 3/06) of 28

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code FT0348

2. Project title

Sustainable Waste Management in the Chilled Food Sector

3. Contractororganisation(s)

Imperial College London(Centre for Environmental Policy,Applied Economics and BusinessManagement Research Section)

54. Total Defra project costs £ 54,628(agreed fixed price)

5. Project: start date................ 01 July 2004

end date................. 31 August 2007

SID 5 (Rev. 3/06) of 28

6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They

should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the

intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.Within the UK food industry, the chilled food sector is undergoing rapid growth. Chilled foods have short shelf lives and are susceptible to spoilage if not used within the specified time. Most chilled food manufacturers buy in ready prepared ingredients, which also have short shelf lives. These trends have implications for the quantity of food and packaging waste generated. The objectives of this study were to ascertain how much waste occurs in the chilled food supply chain, where and why it occurs, and how it might be reduced. Emphasis was placed on generic issues rather than on those related to the manufacture of specific types of product.

The identification of opportunities for waste reduction requires an analysis of current activities and the waste arising from them. Value Stream Mapping (VSM) is a diagnostic technique that originated in Lean Manufacturing for the purpose of eliminating wasteful activities and reducing production lead-time. This research has extended Sustainable Value Stream Mapping, a recently developed method that incorporates the VSM approach, to include a suite of environmental parameters in addition to operational measures in order to evaluate the waste generated in the value stream of a chilled food product.

Anecdotal evidence from one-day visits to six chilled food manufacturers implied that volatility in retailers’ order quantities coupled with forecast inaccuracy make it difficult for manufacturers to estimate material requirements and to plan production, thus reducing efficiency and encouraging over-production to ensure availability, factors that increase both physical and operational wastes. Detailed case studies were carried out on various value streams to assess the wastes derived from discrete activities and to analyse the underlying causes. Particular emphasis was placed on the supplier-retailer interface, as this relationship has important implications for the entire supply chain. The retailers’ perspective on waste was ascertained from a meeting with representatives from the British Retail Consortium (BRC) and from conference presentations by senior managers from two major supermarket chains.

In the opinion of the BRC, waste at retail outlets is driven by food safety concerns, legislative requirements, marketing, poor vendor compliance and consumer expectations of constant availability and uniformity of appearance, which is often mistaken for higher quality. The major issue for retailers is to ensure availability to prevent a loss of customer loyalty. This can cause over-supply and waste. The views expressed in conference presentations by senior managers from two UK supermarket chains provided insights into their relationships with suppliers. Collaborative partnerships are helpful for improving demand management and reducing inventories but are largely reserved for the major branded suppliers, not the SMEs producing own-label items. Suppliers are expected to turn the data provided on I.T. systems into “useful information”. This can prove difficult for the small companies with low profit

SID 5 (Rev. 3/06) of 28

margins and limited resources that are typically own-label suppliers. The responsibility for “getting it right” is pushed back onto the suppliers, who must bear the risks from holding buffer stocks to ensure availability with the risk of non-requirement and, in the case of chilled foods, the economic and environmental costs attributable to refrigerated storage.

Manufacturers and suppliers of chilled food products to the major supermarket chains face a number of generic issues. Order lead-times are generally less than 24 hours but production (or supply) lead-times can be several days, so production has to start in advance of receiving final orders, with quantities based on forecasts that are inevitably inaccurate. Under-estimates can cause difficulties in planning production schedules for high numbers of product items because additional short production runs must be slotted in to meet shortfalls, and shorter runs cause more changeover waste in terms of both physical wastes and production time. To ensure availability in response to retailers’ volatile ordering patterns and to reduce production-related waste from short runs, suppliers hold buffer stocks in spite of the associated risks and costs. Key factors affecting waste at the chilled food manufacturer are the type of product (whether cooked or uncooked); length of shelf life (cooked products have longer shelf lives); and length of order lead-time versus production lead-time (fresh produce and cooked products have longer supply/production lead-times than uncooked products made from bought in, ready prepared ingredients).

The findings on wastes and greenhouse gas emissions from two case studies were scaled to UK level. For UK imports of Spanish tomatoes (all varieties) totalling 190,000 tonnes annually, a similar level of waste to that in the case study would cause about 14,700 tonnes of wasted tomatoes within the supply chain before reaching the end-consumer. About 12,200 tonnes of cardboard would be required to import the total quantity, of which around 940 tonnes would have arisen from the quantity wasted in the supply chain. Approximately 7,000 tonnes of CO2 equivalents per year could be attributed to the life cycles of these wasted tomatoes (if landfilled) and to the cardboard packaging arising from those wastes (if recycled). For UK sales of all brands of pastry products in the same varieties as those analysed in another case study and amounting to around 12,900 tonnes annually, assuming similar levels of waste to the case study production of these products would cause at the manufacturing stage only approximately 2,900 tonnes of food waste, 40 tonnes of plastic waste, 35 tonnes of cardboard and paper waste, and 9 tonnes of metal wastes, and would incur 70 million litres of water consumption and 5,400 tonnes of greenhouse gas emissions from energy use per year. The quantity of food waste would be closer to 1,500 tonnes per year if the percentage calculated as being due to equipment-related pastry trimmings in the case study were excluded. These figures do not include wastes and emissions at the other life cycle stages of raw material production, primary processing, use, disposal, or the intervening transport stages. A reduction in the requirement for chilled storage areas by all the UK’s chilled food manufacturers by holding lower levels of buffer stocks was estimated to potentially save, at the very least, tens of thousands of tonnes per year of greenhouse gas emissions from energy consumption but this would require greater certainty in order quantities.

From a survey of chilled food manufacturers, machine waste from set-up, changeover and cleaning had a median value of 8% of total food waste. A certain proportion of this will be caused by additional changeovers due to discrepancies between forecasts and actual orders. Total out of life wastes had a median value of 27% of total food waste. Wasted out of life intermediate and finished products both had medians of 5% of total food waste and, as value-added items, will have incurred significant energy consumption from cooking processes (if required) and refrigerated storage. The opinions of the survey respondents were consistent with the case study findings that such wastes could be reduced if retailers made more data available, ensured that the data was reliable, made fewer late changes to order quantities, and were more open to dialogue with smaller suppliers. An extension of order lead-time appropriate to the shelf life of the product and to the production lead-time would allow chilled food manufacturers to produce to demand rather than to an inaccurate forecast. Producing to demand would also require suppliers to improve their efficiency by cutting lead-times where possible, perhaps by reviewing the length of time spent by intermediate products in refrigerated storage.

The opportunities for chilled food manufacturers to reduce waste are constrained by real-world difficulties. The internal constraints are the time required to identify and implement the necessary changes, the practicability of these changes, costs vs. benefits, expertise and resources. The external constraints, which are likely to be more difficult to overcome, are supply chain relationships, particularly with regard to the dominant position and current procedures of retailers, and the attitudes and expectations of consumers in relation to on-shelf availability, uniformity of appearance as a misleading indicator of quality and willingness to accept substitutes.

It should be noted that these findings are based on only a small amount of data from two case studies and twelve survey responses but are substantiated by anecdotal evidence from site visits to eight other chilled food manufacturers.

SID 5 (Rev. 3/06) of 28

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with

details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the scientific objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Transfer).

1. Research Objectives

1.1 Objectives as Set Out in the ContractThe main purpose of this research was to analyse waste in the chilled food supply chain in order to answer four basic questions: how much waste is generated, where in the supply chain does it arise, why does it occur, and how might it be reduced? In order to assess the sustainability of current practices, the investigation of both economic and environmental factors was essential. An additional purpose of the research was to develop a methodology for analysing and reducing physical wastes that would be feasible for use by SMEs with limited resources, the principal motivation for which would be the associated reductions in costs. The method should preferably be understood by all levels of staff, from shop floor to senior management, because those working within an industry have a greater understanding of the day-to-day problems that cause waste and, with raised awareness, are better able to resolve them.

1.2 Extent to Which Objectives Have Been MetWhilst it has not been possible to generate sufficient data in order to make estimates of waste generated in the chilled food industry as a whole, there was sufficient and consistent evidence that poor demand management (volatile orders, late changes to orders and inaccurate forecasts) is a significant source of inefficiency and physical waste across the chilled food industry. Whilst the waste generated at the interface between retailers and manufacturers might be small compared to that which arises upstream in agricultural production and primary processing, the relatively low costs associated with its reduction make it a ‘low hanging fruit’ in the battle against waste and the quest for sustainable food chains. Moreover, volatility downstream invariably generates even greater volatility upstream. Thus, whilst this research did not seek to measure it, the relationship between volatility and inefficiency is likely to be amplified further still upstream, so any improvements downstream are likely to benefit the food chain as a whole.

The Sustainable Value Stream Mapping methodology developed during the course of the research has considerable merit, not least the extent to which it should be possible for small food manufacturers to use it themselves, without the support of external consultants. However, the paucity of information relating to energy use in many small businesses means that in most cases estimates have to be made which can obviously reduce the validity and reliability of the survey instrument.

2. Research Methodology and Strategy

2.1 IntroductionThe identification of potential opportunities for waste reduction in the supply chain requires an analysis of the quantities and types of waste arising from individual processes at the product level. Numerous product-oriented environmental analysis methods have arisen from the natural sciences, the most widely accepted of which is Life Cycle Assessment, but such methods do not specifically address factors relating to economics or the marketplace. An alternative approach to waste reduction that originated in the automotive industry is Lean Manufacturing, the principal aim of which is to improve value for the customer by eliminating wasteful activities

SID 5 (Rev. 3/06) of 28

from the production process, thereby reducing costs. The goal of Lean Thinking, derived from Lean Manufacturing, is to promote team working and provide the insights required to recognize and eliminate waste. Although the Lean approach aims to identify wasted resources, its emphasis is on operational aspects rather than on environmental burdens. However, Sustainable Value Stream Mapping is a recently developed method that is underpinned by Lean principles and addresses both operational and environmental factors.

2.2 The Lean ParadigmTaiichi Ohno formulated the ideas behind what is now known as Lean Manufacturing. His simple philosophy was to reduce the time-line between the receipt of an order and the receipt of payment for its delivery by eliminating the non-value-adding wastes that he called muda, frequently called the Seven Wastes, thereby improving efficiency (Ohno, 1988). The Seven Wastes are:

1) Waiting, by operators and machines;2) Transportation of materials;3) Unnecessary or overcomplicated processes;4) Excess stock or materials (inventory);5) Excess movement by operators;6) Defective products;7) Overproduction.

The Seven Wastes lead to low productivity, poor quality and increased costs. Ohno’s ideas for reducing muda formed the basis for Lean Thinking (Womack and Jones, 2003).

2.3 Basic Principles of Lean ThinkingLean Thinking has five basic principles (Womack and Jones, 2003).

1) Specify valueDelivering value is the essence of Lean Thinking. Value is considered to be what the end-consumer is willing to pay for, and it can only be expressed in reference to a specific product that meets the consumer’s needs at a specific price and at a specific time. All participants in the value stream must grasp the idea that only the end-consumer defines value, so they should work towards delivering more efficiently whatever that single definition encompasses.

2) Identify the value streamThe value stream comprises all the activities in the creation of a specific product, from the procurement of raw materials up to the point of sale to the end-consumer. Analysis of the value stream typically identifies three types of activity: those that create value, those that create no value but are essential, and those that create no value and are avoidable. The avoidable activities that create no value are the wasteful steps that need to be eliminated.

3) Make value flowAfter eliminating wasteful activities, the remaining activities that create value are made more efficient by introducing greater flow. Lean proponents argue that the product should be worked on without interruptions in the production chain so that it is completed more quickly and with less likelihood of defects or damage, thereby reducing inventory levels, production lead-times and costs. This contrasts with the ‘batch and queue’ methods used in supply led mass production that typically cause excess inventories at various stages.

4) Let the customer pull valueThis is another way of saying that production should be demand led. The reduction in lead times achieved by improving flow allows production to be more responsive to actual demand and less reliant on forecasts.

5) Pursue perfectionOnce the value stream has undergone a Lean transformation, further reductions in wasteful activities become apparent in a process of continuous improvement that aims to deliver increasing value to the end-consumer.

Jones and Womack (2003) advise that three wastes in particular, overproduction, unnecessary inventories and unnecessary transportation, should be the primary focus for reduction and that this should be done by improving information flows and logistics. They suggest that overproduction at process and facility level is caused by poor intra-facility information flows and the requirement for meeting targets for equipment utilization. At the value stream level, they consider unnecessary inventories to be due to “erratic information flows between firms and facilities as well as incapable and batch-oriented upstream processes”, and unnecessary transportation to be due to “location decisions that seek to optimize performance at individual points along the value stream rather than the whole value stream”. Consequently, Jones and Womack (2003) specify the following requirements for a lean value stream.

All firms in the value stream need to be aware of the rate of demand for the product by the end-consumer so that real demand, the ‘signal’, can be distinguished from ‘noise’ caused by distortions in demand, such as the bullwhip effect.

SID 5 (Rev. 3/06) of 28

There should be the minimum level of inventory (raw materials, work in progress and finished goods) required to support demand from the next process downstream.

There should be as few transport links as possible, with the elimination of links being preferable to their speeding up by, for example, air-freighting.

Information flows should comprise “pure signal and no noise”. The lead-time should be as short as possible in order to respond to real demand rather than forecasts. Any changes implemented to bring about the above improvements should be at minimum or zero cost, with

the easiest and quickest changes carried out before incurring capital outlay.

2.4 Value Stream MappingValue stream mapping (VSM) is a diagnostic tool for visualizing the value stream for a specific product or product family (Rother and Shook, 2003). The first step is to draw a current-state map by ‘walking through’ a specific product’s value stream door to door within a plant. Using the information gathered, a desired future state is developed and mapped. Figure 2.1 shows a general example of a value stream map. The basic processes in the manufacture of the product are drawn from left to right across the lower half of the map, so that one process box represents one area of material flow. Data boxes are added below each process box, showing measurements such as cycle time (the time taken to complete a process so as to produce a single item), value-adding time (the time spent on the essential activities required to create the product) and changeover time (the time taken to switch from one type of product to another). Across the upper half of the map, information flows (indicated by arrows) are recorded along with their frequency, starting on the right with the customer, via the plant’s Material Requirements Planning (MRP) or production control system, and ending on the left with the raw materials supplier(s). Information flows are a crucial aspect of VSM as they are the drivers for events at shop-floor level, including the generation of waste.

Figure 2.1 General example of a value stream map

Source: Norton (2007). Drawn using eVSM v.2.3 (GumshoeKI, Inc. 2000-2005).

2.5 Sustainable Value Stream MappingClearly VSM does not explicitly consider environmental performance, which may or may not be improved by a Lean implementation. Therefore Simons and Mason (2002) have proposed a novel method called Sustainable Value Stream Mapping (SVSM) that enhances sustainability in product manufacture by analysing emissions of the greenhouse gas, CO2, in addition to value-adding time. SVSM is intended to be “a simple, do-it-yourself method for establishing the facts” relating to sustainability in the procurement and distribution of products (Simons and Mason, 2003). The aim is to maximize the proportion of value-adding time and minimize carbon dioxide emissions over the supply chain as a whole, as follows:

Maximize “Value Add %” = TimeChain Supply Total100Time Adding ValueChain Supply ×

Minimize “CO2 %” = Product ofght Market Wei100COChain Supply 2×

SID 5 (Rev. 3/06) of 28

Any economic and environmental benefits thus achieved are assumed to be accompanied by social benefits and therefore contribute to sustainability (Simons and Mason, 2002).

A map is constructed in a similar way to conventional value stream mapping, categorizing activities as value-adding (VA), necessary but non-value-adding (NNVA) and non-value-adding (NVA), but for each process activity and transport activity, both the VA/NNVA/NVA time and CO2 emissions are quantified.

Mason et al. (2002) used SVSM in a study for the Department of Transport that modelled CO2 emissions from farm gate to retail outlet for alternative distribution scenarios for three supply chains: lettuce, apples and cherries. However, only the CO2 emissions from the transport activities were quantified; process activities were excluded.

2.6 Extending SVSM to Include Other Environmental Performance IndicatorsIn the USA, the Environmental Protection Agency advises integrating Lean implementation with the achievement of environmental performance goals (Anon., 2006a). They describe environmental wastes as “any unnecessary use of resources, or substance released to the air, water or land that could harm human health or the environment” and emphasize that “environmental wastes, although not considered one of Lean’s seven deadly wastes, are embedded in or related to the wastes targeted by Lean methods”. In other words, environmental and operational wastes are inextricably linked.

One of the purposes of this research was to further extend SVSM beyond the measurement of value-adding time and CO2 and to include additional environmental performance indicators (EPIs), in particular to carry out an evaluation of environmental wastes over a product’s value stream in order to analyse the association of those wastes with discrete activities. This method was considered a pragmatic approach because both material and information flows are investigated, allowing waste generation and its underlying causes to be analysed.

DEFRA (2001) recommends the use of certain EPIs to help companies describe their current most important impacts and to form a basis for improvement targets. The Food and Drink Federation (FDF) have also published guidelines for the UK food and drink manufacturing industries in which Key Performance Indicators are proposed (Anon., 2002a). The EPIs recommended by DEFRA and the FDF provide a basis for the analysis of environmental wastes and resource usage in the chilled food sector using SVSM. Those selected for this research were:

Food wastes; Packaging wastes (plastics, cardboard, paper and metals); Energy consumption and the associated greenhouse gas emissions; Water consumption.

2.7 Conceptual Framework

Figure 2.2 Conceptual Framework

Source: Norton (2007)

SID 5 (Rev. 3/06) of 28

Figure 2.2 shows a conceptual framework applicable to both VSM and SVSM that illustrates the ideal sequence of events. The framework shows that the measurement of performance indicators for activities in the current state raises awareness and helps to identify the processes that need to be changed. The value stream map thus created serves as a basis for planning the changes that, when implemented, will lead to the desired future state. Changes in behaviour are essential in order to achieve the future state, which ideally involves all supply chain partners. The future state becomes the next current state in a process of continuous improvement that identifies further waste for elimination.

2.8 Research StrategyThere are several different types of research strategy, such as an experiment, a survey, archival analysis, a history or a case study (Yin, 1994). In the context of this research, two of the questions to be addressed were “how much waste occurs in the supply chain of a particular product and where”. This type of data is not in the public domain, ruling out an archival analysis and leaving a survey as a possible option. But additional questions were posed: “why does waste occur and how might it be reduced”. Such questions require the use of experiments, histories or case studies, because these strategies are of an explanatory nature, whereas surveys are more concerned with incidence or prevalence, and are limited by the number of questions that can be asked (Yin, 1994). An experiment has to be ruled out because it requires control over behavioural events and a history, by definition, is only applicable to past events. The only remaining strategy for this research was a case study approach.

The major unit of analysis was the chilled food sector of the UK food industry. As there are numerous types of chilled food product, an investigation of as many different types as was practicable would bring benefits in terms of analysis of the issues within and across different product groups. This required multiple units of analysis within the context of the chilled food sector, leading to what Yin (1994) describes as a multiple-case design. Replication logic may then be followed; that is, if certain causes of waste were the same in all cases then replication would have occurred, allowing a theory to be developed prompting further research as to general prevalence. It should be noted that replication logic is not the same as sampling logic where, from the outset, a number of subjects are assumed to be representative of the entire population, allowing statistical procedures to be applied and confidence intervals established (Yin, 1994).

2.9 Selection of Case StudiesSix companies initially agreed to participate in the project:

A supplier of whole salad products; A manufacturer of processed meat products; A manufacturer of pastry products; Two ready meals manufacturers; A company supplying and replenishing automatic vending machines, in which items sold included chilled

foods.

For various reasons, some companies subsequently withdrew support, and ultimately the only companies from the original six where meaningful amounts of data were obtained were the supplier of salad products, the manufacturer of pastry products and the vending machine supplier. At no individual site was access gained to all the data that would have been useful in ideal circumstances. Any missing or non-specific data had to be replaced or attributed by assumptions and estimates based on observing certain activities for a limited period, informed opinions from senior staff within the companies, and the apportionment and allocation of historical data. In all cases confidentiality had to be assured, given the commercially sensitive nature of the investigation.

2.10 Data CollectionAt each participating company, as many as possible of the selected EPIs were measured and attributed to discrete activities. Where attribution to specific activities was impractical because of multifunctionality, values were allocated using methods similar to those recommended in life cycle assessment (LCA) (Guinee, 2002). Each activity comprises numerous individual work elements, and all such elements would normally be measured when carrying out standard VSM (Rother and Harris, 2001). However, a top-down approach was considered more appropriate for this project: firstly, because many activities in food production are automated, preventing their division into separate elements; and, secondly, because the aim was to capture the ‘big picture’. This top-down approach involved deciding whether each activity was predominantly VA, NVA or NNVA and recording it as such. Also, this research was more concerned with the strategic and commercial issues that cause waste because these were more likely to be generically applicable across the chilled food sector as a whole, allowing the use of replication logic regardless of product. Process-related issues at the shop-floor level, whilst not uninteresting, were likely to be generically applicable only to those companies producing similar products by similar methods, restricting their usefulness in terms of replication logic on a sector-wide basis.

SID 5 (Rev. 3/06) of 28

Mapping environmental wastes during a ‘one-off’ exercise was useful for identifying types of waste and where they arose but would provide only a snapshot of the quantities involved, with no way of knowing whether they were typical. Therefore all participating companies were asked to provide copies of any historical waste data they had in as much detail and covering as long a period as practicable in order to capture seasonal variations. Details of forecast quantities and actual order quantities for the period covered by the waste data were also sought. Data relating to electricity, gas and water consumption were requested in as much detail and for as long a period as could be provided. Additional understanding of the production and waste issues within the company and of the relationships with customers and suppliers was obtained by semi-structured interviews with key staff.

3. Preliminary Indications of the Causes of Waste

3.1 Observational Visits to Chilled Food ManufacturersIn 2005, the Process Industry Centre for Manufacturing Excellence (PICME) and the Food Processing Faraday Partnership (FPFP) undertook a DEFRA-funded waste minimization study of eight UK chilled food manufacturing sites (Anon., 2006b). The author participated in this study as an observer, visiting six sites alongside the process engineers from PICME and a food technologist from FPFP. Visits to two further companies that had initially agreed to be case studies for the main research project but subsequently withdrew their cooperation provided additional anecdotal evidence.

The eight sites visited during this preliminary phase covered a wide range of chilled food products and included:

Two sandwich manufacturers; Five manufacturers of convenience foods and ready meals (pasta-based, rice-based and potato-topped); A manufacturer of pastry products.

Taking all the visits together, some important generic issues were apparent.

Production planning was based on forecasts, which were inevitably inaccurate and frequently led to over-production and potentially high levels of waste. This was particularly pertinent to products that required cooking, where lead times were longer than for uncooked products, such as sandwiches, and where production started far in advance of receiving the actual order.

Nearly all companies bought in many of their ingredients ready prepared so most trimming waste occurred upstream in the supply chain. These ready-prepared items also have a short shelf life (often only 2 days if not frozen) so inaccurate forecasting for these can also lead to waste.

All eight companies alluded to increased waste levels for food and/or packaging resulting from the actions of retailers, with the main reasons being orders that differed markedly from forecasts (whether supplier- or retailer-generated), high variability in order quantities, short order lead-times, frequent changes to packaging design, short-term changes to packaging for promotions, and issues of quality and/or aesthetics.

3.2 Views of the British Retail ConsortiumIn 2006, the British Retail Consortium (BRC) agreed to a meeting with the author to discuss the retail perspective. The Head of Technical Services and the Assistant Director of Food Policy gave their views on the causes of waste at food retail outlets and on the allegations by suppliers that retailers exacerbate waste in the food supply chain. They said that retailers did not want to waste food, as to them it represents a cost they would prefer to avoid. Several reasons were put forward for the unavoidable disposal of food at retail outlets (Martinez-Inchausti, 2006).

1. Items out of shelf life have to be discarded, even though they may be perfectly safe. The Use By date has to reflect the fact that the consumer may not handle and/or store chilled items appropriately, so there has to be a safety margin.

2. Production problems may make withdrawals necessary because of microbiological issues that occurred during manufacture.

3. Residues in the food may exceed established parameters. Withdrawals may occur because, for example, the maximum limits for pesticides, heavy metals, mycotoxins or microbiological factors were exceeded.

4. Changes in legislation might give rise to a change in composition leading to withdrawals over a transition period, although this is not relevant to chilled foods because of the short shelf life.

5. New safety limits may come into effect for certain ingredients.6. Labelling may contain inaccurate information about the contents.7. There may have been errors during manufacture. For example, the product may contain trace levels of

allergens not usually found in the product and not declared on the label.

SID 5 (Rev. 3/06) of 28

8. Harmonization across EU member states with different cultures and cuisine can lead to more waste, even though not always relevant to UK cooking and eating habits.

The disposal of obsolete packaging may also be unavoidable for the following reasons.

1. Legal reasons, such as changes in labelling laws concerning, for example, allergens, genetically modified ingredients, health marks, or new laws relating to food contact materials.

2. Marketing reasons, such as promotions, changed recipes, or changes in the appearance of packaging.

The opinion of the BRC representatives was that marketing now caused only a very small proportion of wasted packaging upstream at the food manufacturer. Historically, designs might have been changed on a whim but now packs were usually re-designed to coincide with changes in legislative requirements, and usually over a period during which stock ought to be used up. Surplus packaging waste at the food manufacturer might be caused by minimum print run sizes.

Other issues regarding waste were raised, and the following explanations for retailers’ actions were given (Swoffer, 2006).

Some waste was the result of “poor suppliers and poor sourcing”. For example, a retailer might visit a fresh produce pack house and find that products are below specification. The products might have to be re-worked, causing wasted packaging only, or all the packed items might be thrown away. The BRC felt that this is related to the quality of organization at the supplier level, and that it should be controllable.

Marketing standards set by DEFRA, that is, Class I, Class II, and so on, are legal requirements. For example, undersized or mis-shaped items are unacceptable in Class I. This system dictates uniformity and UK consumers now expect uniformity. Store managers might have encouraged this to some extent because they always placed fresh produce near the store entrance so that displays looked aesthetically appealing, hence the desire for uniformity in size, shape and colour.

Retailers expect deliveries to be within specification, on time and in full. For example, if sandwiches are not delivered during the morning, they will not be sold at lunchtime and will have to be thrown away, so late deliveries of items with a limited opportunity for sale might be rejected in their entirety.

The unit size supplied, that is, the number of packs per case/tray delivered, can cause inefficiency and increase costs. The fewer packs per case/tray, the more handling that is required and the higher the packaging and distribution costs. However, even when only a small number of particular items are sold, the retailer might lose custom if they are unavailable.

Price reductions on items nearing the end of their shelf life usually occur on the day before the Use By date. Store operatives have to re-visit shelves several times to continually reduce the price until the item is either sold or expired. The cost of staff time is greater than the money made on the reduced items. For example, when the Head of Technical Services worked for a supermarket chain, they estimated that marking down such items cost £11 million per year in labour and lost margin.

The most important issue was on-shelf availability, and the worst outcome was to have too few products on the shelf, leading to a fall in customer numbers. Before EPOS (electronic point of sale), store managers always over-ordered to ensure availability but EPOS now drives orders by making use of intelligent systems. This should prevent excessive over-ordering. However, the system can be over-ridden on key lines. For example, there may have been poor sales in the week prior to a Bank Holiday but good weather is forecast for the holiday weekend, so the automatically generated low estimates on the system are over-ridden and order sizes increased on specific items, such as barbecue food.

In summary, the BRC representatives considered that many of the actions of retailers are largely driven by the high expectations of consumers, with which they have to comply or else lose business.

3.3 Views of Major RetailersMost of the manufacturers visited had raised the issues of inaccurate forecasting and order volatility as important drivers of waste and inefficiency. The major food retailers were approached to discuss forecasting and ordering, via an invitation distributed by the BRC on behalf of the research team. No replies were received. Therefore it was necessary to gauge opinions indirectly from conference presentations given by senior managers of food retailers. In 2006, the Institute of Grocery Distribution (IGD) hosted a conference on Demand Planning and Forecasting. Two of the presenters were from major supermarket chains and relevant aspects of their presentations are summarized below.

SID 5 (Rev. 3/06) of 28

Joint Presentation by Colgate-Palmolive and Tesco (Brierley and Spring, 2006) This presentation was concerned with collaborative planning. The presenters highlighted on-shelf availability as a major issue and argued that this was improved by supplier-retailer collaborations, which they described as “working together to satisfy consumer wishes better, faster and at less cost”. Tesco had the following advice for suppliers wishing to enter into a collaborative relationship:

“Suppliers have to EARN the right to be part of our collaborative process Get the basics right Start by delivering what we order Share working practices and understand how each other works Implant people and share information intensively to adapt practices Develop joint plans and common goals Conduct joint learning activities”.

Tesco meets regularly with collaborative suppliers to produce a joint forecast of likely sales volumes, especially during promotions. The Tesco presenter pointed out that Tesco made daily EPOS data available to all suppliers but questioned whether some suppliers understood it and knew what to do with it. A member of the audience asked whether small suppliers with a turnover of, say, £20-25 million were likely to be able to enter collaborative relationships and have regular planning meetings with Tesco in the same way that Colgate-Palmolive did. The Teso presenter replied that it was unlikely, but that they could always “make a phone call” to Tesco. The Colgate-Palmolive presenter said that their UK operation was small with few employees, inferring that they had no more clout than a small company. There was no acknowledgement by either presenter that small, private companies were unlikely to have the expertise or resources of a multinational corporation to carry out a sophisticated analysis of EPOS data.

Presentation by ASDA (Ellis, 2006)The Supply Chain Director of ASDA gave a presentation on retailer demand. In ASDA’s view, “the customer is king” and the goods required by the consumer should be available on the shelf, not at the back of the store or elsewhere within the supply chain. He said that an important factor affecting availability was vendor compliance. He also considered that high levels of inventory were not a guarantee of good availability and that the key was to get stock levels right, requiring joint decision-making by ASDA and their suppliers. To encourage this, there were two computer systems that ASDA made available to all their suppliers, Retail Link and CPFR (Collaborative Planning, Forecasting and Replenishment). However, he said that the data from these systems had to be “turned into useful information”. So, once again, the retailer was passing responsibility back to the supplier. However, his view was that “lots of noise in forecasts is due to activity in the system” and “created by the industry itself”. Therefore ASDA would prefer no promotions, because “availability and loyalty were more important than uplift from short-term promotions”.

These two presentations highlight an important issue that warrants further research but was beyond the scope of this project; supermarkets clearly believe that most of the waste problems in the food chain are not of their making but are due to the failure of suppliers to comply with their requests and to make use of the information they provide. However, empirical evidence from the academic literature (for example, Taylor and Fearne, 2006) and anecdotal evidence from food manufacturers, the trade press and personal communications suggest that there is inconsistency in the implementation of collaborative forecasting, demand management and order processing. The best practice that is claimed and promoted through ECR UK, which invariably involves the larger branded manufacturers, bears little resemblance to the way in which retailers communicate, and their systems integrate, with the smaller and own-label suppliers that are typical of the chilled food sector.

4. Case Studies

Following the preliminary familiarization phase, in-depth case studies were carried out at three companies supplying different types of product: salad products, with cherry tomatoes being the focus of study; pastry products; and the automatic vending of chilled foods. Findings from these case studies are set out in Sections 4.1 - 4.3. Value stream maps and tables of detailed data for are provided in Appendix 1.

4.1 Case Study 1: Fresh Salad Products (Cherry Tomatoes)For reasons of confidentiality, neither the supplier nor its customer will be identified, but for ease of reference the supplier will be called Case A.

4.1.1 Background to the Company and the Waste ProblemCase A sources and packs whole salad products for its exclusive customer, a supermarket chain. When the customer asked Case A to improve quality, costs for Case A were greatly increased, as they had to recruit more quality control staff and significantly more waste was generated due to an increased rejection rate. Any deliveries

SID 5 (Rev. 3/06) of 28

rejected by the retailer are returned to Case A. Some rejects, if satisfactory, are sold to wholesale markets at the highly variable, daily spot price otherwise they are dumped.

Fresh produce has a short functional life and undergoes progressive deterioration down the supply chain, especially if cool chain integrity is not maintained appropriately. Journeys from Spain take up to three days by road but Spanish drivers do not set out on Sundays, as harvesting and despatch are not carried out. Deliveries initiated on Monday do not arrive until Wednesday but the products have to be in store for the weekend. Demand for salad products is highly variable, with peaks occurring around public holidays and during periods of fine weather. So, there is considerable uncertainty in the supply chain, caused by a combination of volatile demand and volatile supply, making some level of waste inevitable.

4.1.2 Summary of Case Study FindingsRoot Causes of Waste The predominant causes of waste and wasted resources in the cherry tomato value stream analysed in the case study are:

the quality of raw materials, especially of non-Spanish imports; the long lead-time for supply compared with the short order lead-time; the holding of buffer stocks for an average of three days in refrigerated conditions to ensure

a) temperature compliance, andb) availability, to compensate for forecast inaccuracy and order volatility.

The first two causes are external to the organization. The third cause, although largely external, is exacerbated by internal practices at Case A.

The most probable cause of waste at the retailer is over-supply. To a large extent, this is likely to be deliberate in order to ensure availability. Also, a small proportion is due to taking extra from Case A when they are over-stocked. However, when spread across a large number of branches, the total quantity of waste per branch amounts to only a few cases per SKU per week, and one might question whether it is possible to be more accurate whilst trying to ensure availability.

Quantities of Wastes and EmissionsOf the total quantity of cherry tomatoes imported by Case A, 10% were wasted, 5.5% at Case A and 4.5% at retail outlets. As shown in Table 4.1, Spanish produce made up an estimated 80% of imports but gave rise to slightly less than half the tomato waste (or 3.4% of Spanish imports), with the remaining 20% of imports from Morocco and Israel giving rise to the remaining half of the waste (or 14.1% of non-Spanish imports). In the Spanish tomato supply chain analysed, 229 tonnes of cherry tomatoes were wasted at Case A and 296 tonnes at retail outlets during a 12-month period.

Table 4.1 Percentage of cherry tomato imports wasted at Case A, by origin, during a 12-month period

Estimate of Quantity Imported (kg)

Total Quantity Wasted (kg)

Waste as % of Quantity Imported

Imports from Spain 6,800,715 229,147 3.4

Imports from Morocco and Israel 1,700,179 239,457 14.1

All Imports 8,500,894 468,604 5.5


A total of 435 tonnes of cardboard cases were used to import the Spanish tomatoes, of which 36 tonnes can be attributed to the import of wasted produce. The total quantity of plastic packaging (PET and PP) used for retail sale and requiring disposal was 194 tonnes, of which 9 tonnes arose from unsold items at the retail outlet.

Greenhouse gas emissions were roughly estimated for the life cycles of the wasted Spanish cherry tomatoes and wasted packaging, although this was not possible for some stages of the life cycle therefore values calculated are minima (see Appendix 1). In the case study, a total of 426 tonnes of CO2 equivalents were attributable to wasted

SID 5 (Rev. 3/06) of 28

tomatoes and to the packaging arising from those wastes over a 12-month period. This total is reduced to 368 tonnes if wasted tomatoes are composted in preference to disposal at landfill.

4.2 Case Study 2: Pastry ProductsIn contrast to Case A where simple, unprocessed products were sourced and packed, the next case study investigated the manufacture of complex chilled products. In order to maintain confidentiality, the company will be called Case B.

4.2.1 Background to the Company and the Waste ProblemCase B produces pastry products, which are sold to several supermarket chains. Production levels are around 10,000 tonnes per year comprised of a high number of SKUs, many of which are low volume products made in a batch production process.

A significant quantity of food waste arises from sub-standard items. Most of the remainder comprises pastry trimmings that cannot be re-used in case of cross-contamination between different varieties of product. Uncontaminated pastry is re-used whenever possible but reduces the quality if re-worked excessively. Of the average 49 tonnes of food waste each week, over 40 tonnes is unavoidable pastry waste that is largely related to the nature of the product and the production process. The company recognizes that it would be possible to reduce this pastry waste by buying alternative production machinery but the capital outlay makes this a difficult commercial decision because, in common with the rest of the chilled food sector, the company attains only low profit margins. (In fact, since the case study was carried out, the company has purchased this alternative equipment, which has reduced the quantity of wasted pastry trimmings, but the following findings relate to the period preceding that purchase).

The company has access to the retail sales data of some of their customers, which helps in planning production, but the quality of I.T. systems and available data varies between retailers. Volatility in order quantities combined with short order lead-times can disrupt production schedules so that the most efficient production sequence cannot always be followed, prompting additional changeovers to make up for shortfalls if order quantities exceed forecasts. Conversely, if forecasts greatly exceed order quantities, there is a risk of waste from over-production.

4.2.2 Summary of Case Study FindingsRoot CausesThe internal causes of waste and wasted resources in the pastry product value stream served by Case B are predominantly due to the nature of the product and the production process, namely:

The high quantity of pastry trimmings arising from equipment that would be expensive to replace; Rejects on production lines due mainly to damage or quality/aesthetic issues; Rejected finished products caused mainly by baking issues that affect appearance, QA/QC procedures,

lightweights and mis-shapes.

The high levels of product-related waste can be compounded when short production runs are necessitated by retailers’ actions, in particular:

Order lead-times that are shorter than production lead-times; Volatile order quantities that increase forecast error; Large discrepancies between retailers’ estimates and actual orders; The quality and quantity of EPOS data on retailers’ I.T. systems; Manual adjustments and/or insufficient updating on I.T. systems by retailers’ staff; The basis used by some retailers for determining order quantities; Unwillingness on the part of some retailers to enter into helpful discussions on estimates and order quantities; Lack of communication on weekends and Bank Holidays.

In order to reduce product-related waste, Case B are obliged to hold buffer stocks at all stages of production and to bear the full risk of waste from out of life items in addition to the energy costs of chilled storage for the following reasons:

1) To facilitate the scheduling of a large number of complex SKUs across a limited amount of production equipment;

2) To improve production efficiency and reduce product-related waste by scheduling longer runs with fewer changeovers;

3) To ensure availability so that any unexpectedly large order quantities can be met.

Quantities of Wastes and EmissionsFrom an analysis of historical data provided, total weekly food waste at Case B averaged 24% of production. Finished product rejects were relatively high at 2.3% of production but were exceeded by production line rejects

SID 5 (Rev. 3/06) of 28

at 3.7% of production, although this is mainly pastry because filling is segregated and re-used if possible. However, waste generated within the production area comprised 88% of total food waste (22% of production), far exceeding that from other causes. Of this amount, the production process itself generates 61% as pastry trimmings, and is purely related to the nature of the product. If this specifically product-related waste is excluded, the remaining waste is closer to 11.5% of production, which is in the range typical for the chilled food sector (Anon., 2006b).

As historical waste data was not available at SKU level, it was necessary to determine typical food waste levels for major activities in the value stream of a specific product family by various means: observation; opinions from supervisory staff; apportionment of historical data; and estimation (see Appendix 1). On averaging the results for the analysed products, each kilogram produced was estimated to incur (at the manufacturing stage only) 226g of food waste, 3.2g of plastic packaging waste, 2.6g of biodegradable packaging waste, 0.7g of metal waste, 5,434g of water consumption and 419g of CO2 emissions from energy use (Table 4.2).

Table 4.2 Average values for CO2 emissions, water consumption and waste at the manufacturing stage per kilogram of finished product based on the analysed family of pastry products at Case B

Energy-relatedCO2 emissions

(g/kg)

WaterConsumption

(g/kg)

FoodWaste(g/kg)

Packaging Waste (g/kg)

Plastics Card/Paper Metals

Product A 471 5461 207 2.9 1.8 0.6Product B 466 5369 226 3.4 2.4 0.5Product C 385 5375 241 2.9 2.1 1.1Product D 355 5531 229 3.4 3.9 0.5AVERAGE 419 5434 226 3.2 2.6 0.7Source: Norton (2007)

Potential Environmental Benefits of Longer Order Lead TimesIf orders were received prior to commencing production, it would in theory lead to a reduction in the requirement for chilled storage areas used for holding buffer stocks and hence a reduction in electricity consumption. At Case B, an extra day on the order lead-time might halve buffer stocks but it is unlikely that the requirement for chilled storage would be halved, as extra capacity is needed for busy periods. However, it might be possible to shut down, say, three medium-sized chill rooms at Case B, reducing the area of chilled storage for intermediate and finished products by 20%, equivalent to 5% of the total temperature-controlled area. The Engineering Manager considered refrigeration to be a major consumer of electricity at Case B, particularly because cooking, baking and water heating were fuelled by gas, but he did not know the proportion of total consumption. The question arises as to how much energy might be saved at the site by reducing the temperature controlled area by 5%. Table 4.3 shows estimates of the potential savings in annual electricity consumption and the resulting CO2 emissions by this 5% reduction for two different percentages of refrigeration-related consumption, 20% and 60%. If refrigeration accounted for 20% of electricity consumption, a 5% reduction in the temperature controlled area might save 26 tonnes of CO2 emissions annually (1% of those from electricity consumption). If refrigeration accounted for 60% of electricity consumption, the saving might be 77 tonnes of CO2 emissions (3% of those from electricity consumption). Goodburn (2007) estimates there are around 150 non-SME chilled food manufacturing sites and hundreds of smaller sites in the UK. If the total number of sites were, say, 500 then a similar 5% reduction in chilled storage at each site might lead to savings in CO2 emissions of between 13,000 and 39,000 tonnes annually. It could be argued that large sites with many chill rooms have greater scope for rationalizing and reducing chilled storage so greater savings might be possible.

Table 4.3 Potential savings in annual CO2 emissions resulting from a 5% reduction in the temperature controlled area at Case B

Electricity Consumption(kWh) % of Consumption CO2 emissions

(tonnes)

Savings in CO2

emissions after 5% reduction in

refrigerated area(tonnes)

6,000 100 2,5003,600 60 1,548 771,200 20 516 26


The Effects of Retailers’ Actions on WasteCase B provided examples of the difficulties they face in trying to accurately estimate how much to produce. The first example concerns the reliability of the data on a retailer’s I.T. system. Table 4.4 shows estimates and orders

SID 5 (Rev. 3/06) of 28

for a two-day period. Looking at products 1 and 2 in the table, estimates and orders were both zero because it had been agreed with the retailer that there would be no orders during this period. However, the retailer’s I.T. system still showed a requirement. Such misleading data may prompt items to be produced by small suppliers who do not generate in-house forecasts and who do not have a helpful dialogue with the customer, as was the case with some other chilled food manufacturers interviewed during the project. Table 4.4 also shows that Case B’s estimates for products 3, 4 and 5, calculated in a spreadsheet using a simple moving average, were closer to the final order quantity than the requirement estimated by the retailer using a more sophisticated I.T. system. However, the manager who planned production at Case B was aware that the retailer’s sales teams were able to manually over-ride the computer system, a fact also mentioned by the British Retail Consortium, perhaps underlining a potential cause of forecast error, especially if inaccurate manual adjustments then became incorporated into subsequent forecast calculations.

Table 4.4 Example of unreliable figures for Case B’s products on a retailer’s I.T. system

Friday Saturday

Case B Estimate (cases)

Requirement on Retailer’s

IT System (cases)

Retailer’s Order

(cases)

Case B Estimate (cases)

Requirement on Retailer’s

IT System (cases)

Retailer’s Order

(cases)

Product 1 0 202 0 0 127 0Product 2 0 302 0 0 166 0Product 3 137 56 118 106 65 96Product 4 58 34 59 37 15 40Product 5 200 135 173 145 101 178Source: Norton (2007)

Another example is concerned with a promotion for the same retailer. A recently introduced range of products was on promotion for the first time. During this promotion, no retailer’s estimates were received for a few days and communication was not possible because it was a Bank Holiday period. Therefore Case B had to calculate their own estimates based on data on the retailer’s I.T. system. Earlier that week, the retailer’s estimates had been around 40% lower than actual orders so, in the absence of any figures from the retailer and with no Saturday delivery, Case B understandably expected high orders for Sunday and Monday. In fact, orders were significantly lower (by more than 90% for some products) and a total of 2,685 kg of intermediate and finished products went out of life (OOL) and were wasted (Table 4.5). This amount is equivalent to 5.5% of average weekly food waste at Case B, or 11.7% when purely product-related pastry trimmings are excluded from the weekly average.

Table 4.5 Orders and estimates over a Bank Holiday period during a promotion at Case B

Last Two Orders

Before Bank Holiday(cases)

Case BSun and Mon

Estimates(cases)

Sun and Mon Orders(cases)

OOL Finished Products

(kg)

OOL Intermediate

Products (kg)

Product A 481/337 850/500 45/45 810 262

Product B 550/413 730/700 40/40 1,142 36

Product C 253/331 360/200 49/128 113 25

Product D 189/168 375/300 98/143 62 92

Product E 218/181 400/110 87/148 0 143

Total 2,127 558

Source: Norton (2007) OOL = out of life

4.3 Case Study 3: Automatic Vending of Chilled FoodsFor reasons of confidentiality, the company will be called Case C.

Case C is a small vending machine company supplying chilled food items to a minority of its customers, in addition to its core business of drinks provision to offices and factories. The company provided a spreadsheet of 30 weeks’ worth of data on food items supplied to and wasted from its vending machines.

SID 5 (Rev. 3/06) of 28

Analysis of the data showed that 11,917 sandwiches and rolls in 66 varieties (amounting to 41% of those supplied) were wasted during the 30-week period, largely due to the short shelf life of only 2 days. Assuming an average weight of 200g per item, this amounts to 2.4 tonnes of food waste (about 80kg per week). Levels of waste were particularly high for some product lines, ranging from 16% to 65% for products where the number of items supplied during the 30-week period exceeded 500.

A similar analysis for pies, pasties and sausage rolls, which have a slightly longer shelf life, showed that 5,112 items of 10 varieties (20% of those supplied) were wasted during the 30-week period. Again assuming an average weight of 200g per item, this amounts to 1.0 tonne of food waste (about 34kg per week). Levels of waste ranged from 14% to 38% on product lines where the number of items supplied exceeded 500 for the period.

The total annual weight of chilled food waste discarded by Case C, based on these waste figures, would be about 6 tonnes. Case C does not attempt any form of sales analysis or forecasting, mainly because of being a very small company with limited resources. The quantities of items required for the replenishment of vending machines are merely guessed by the delivery operators, whom the company considers to be good judges of the level of sales for particular products.

An experiment was carried out to determine whether sales could be predicted more accurately by the Ratio-to-Moving-Average Method (Lapin, 1980). Three sandwich products with high levels of sales and wastage were analysed. Sales figures for thirteen weeks were used to predict the likely level of sales for the following thirteen weeks (Table 4.6). The predicted sales were then compared to the actual sales for that period. The cost of any oversupply was calculated using the cost price of the product. The cost of any undersupply was calculated using the mark-up, that is, the lost profit from not supplying sufficient items.

Table 4.6 Comparison of the cost to Case C of oversupply with the cost of inaccurate forecasting for three sandwich products over a 13-week period

Actual Sales

No of Items

Supplied

Difference Between

Sales and Supply

Cost of Over-

supplyForecast

Sales

Total Difference Between

Sales and Forecast

Total Cost of Under- or

Over-supply

Product 1 379 972 593 £545.56 231 -148 £116.76Product 2 216 659 443 £296.81 253 37 £50.46Product 3 368 640 272 £261.12 330 -38 £89.52Totals 963 2271 1308 £1103.49 814 -149 £256.74Source: Norton (2007)Oversupply of these three items cost Case C £1103.49 for the 13-week period. If forecasting had been used, even though inaccurate, the total cost to the company of either under- or over-supply of those items during the same period would have been £256.74, only 23% of the cost of over-supply. In fact the cost of under-supply might have been far less as substitutes may have been purchased. In terms of physical waste for these three products during the 13-week period, the number of items thrown away in the municipal waste by Case C and subsequently sent to landfill was 1308, which is greater than the number actually sold (963). Assuming an average weight of 200g per sandwich, this amounts to 262kg of food waste. If forecasting had been used, only 37 over-supplied items would have been thrown away, weighing around 7kg, which is a reduction of 97%.

This small case study showed that the use of a relatively straightforward forecasting method would greatly reduce the quantity of waste arising from chilled food items supplied by Case C in its vending machines. If a similar 97% reduction in food waste could be achieved across all items, the annual quantity discarded by Case C would fall from 6 tonnes to about 180kg, or 3.5kg per week. However, in practice this massive reduction is unlikely to be achieved for a number of reasons.

The fairly old design of the vending machines used is such that it is possible to purchase fresher items before older items, some of which inevitably reach their expiry date and have to be discarded.

Case C’s customers insist that the vending machines are fully stocked at the start of each working day and that the full range of items from a selection menu is available so that their employees have a wide choice.

Suppliers impose minimum order quantities, which can be too high for very small companies such as Case C.

5. The Predictability of Retailers’ Orders

5.1 Measures of Forecast ErrorInaccurate forecasting is a source of seemingly unavoidable waste in the food industry; even the very best forecasts will be less than 100% accurate. It is essential to measure the errors in forecasts and analyse why they

SID 5 (Rev. 3/06) of 28

occur so that procedures might be improved (Anon., 2006c). One possible method would be to measure the average discrepancy between the forecast quantity (F) and the actual order quantity (A). This would simply be the mean percent error (MPE) (Chockalingam, 2003).

MPE =

€

∑(F − A) ×100

AN

, where N = number of orders.

Chockalingam (2003) does not consider MPE a reliable measure of forecast error, as it can be skewed by small quantities. This skewness can be avoided by using the Mean Absolute Percent Error (MAPE), which is the most widely used measure of forecast error (Gilliland, 2002; Chockalingam, 2003; Anon., 2006c).

MAPE =

€

∑A − F∑A

×100

Lapide (1998) also considers MAPE to be an adequate measure of forecast error in most cases, but not for volatile markets where there is a high level of demand variation. The variation in historical demand can be analysed using the mean absolute percent variation (MAPV), a measure of the average absolute deviation from the mean demand.

MAPV =

€

∑AN

− A

∑A×100 , where N = number of orders.

If MAPE is greater than MAPV, it shows that the forecast added more variation than was naturally present in the demand data. If MAPE is less than MAPV, it shows that the forecast method has been effective to some extent in predicting the demand variation (Lapide, 1998). An analysis of the percent of variation explained (PVE) can be carried out by comparing MAPE to MAPV. This serves to indicate how much of the natural variation has been “understood” by the forecasting method and therefore provides a measure of forecast accuracy (Lapide, 1998).

PVE =

€

(1− MAPEMAPV

) ×100

5.2 Forecast Error at Case BExamination of the percentage error on Case B’s forecasts for different customers provides a means of determining whether some retailers’ orders are easier to predict than others. Table 5.1 shows the MPEs for Case B’s forecasts for all the pastry products ordered daily by their customers over a 17-week period. The various MPEs correspond with anecdotal evidence from the company concerning which customer’s orders were easier to predict, whether as a result of the customer’s superior I.T. system and openness to discussion (Retailer C) or as a result of more stable ordering patterns (Retailer D). The high MPE of Retailer F was considered by Case B to be related to the ordering procedures of the customer, their poorer I.T. system and over-reaction to variation in demand at store level.

In all cases, MAPE (the more reliable measure of error) is less than MAPV so the forecasts produced by Case B have not added more variation than was inherent in the demand data. Perhaps surprisingly, MAPV shows greater variation in demand from Retailer C and Retailer D than from other retailers, tending to conflict with Case B’s anecdotal evidence, but the high PVEs for these customers show that more of that variation was predictable, allowing a more accurate forecast to be produced, as indicated by low MAPEs. Excluding Retailer A, whose orders were very small, the customer with the lowest PVE and the highest MAPE is Retailer E, suggesting that these orders are the most difficult to predict. However, Retailer E provides the longest order lead-time so this unpredictability has less impact on production efficiency. Retailer F has a higher PVE and a lower MAPE than Retailer E, which seems to indicate that their order quantities are easier to forecast, but Retailer F’s demand variation (MAPV) is far higher than that of Retailer E. This, in conjunction with a higher number of SKUs, higher tonnage and much shorter order lead-time might account for the fact that Retailer F is the customer whose orders are generally the most difficult to forecast and have the greatest potential to disrupt production schedules.

In summary, a senior manager at Case B considered that the analysis in Table 5.1 closely reflected “the pain factor” felt in trying to accurately estimate how much to produce for each customer. High variation in demand is not necessarily problematic if combined with either relatively stable ordering patterns and long order lead-times (Retailer D) or highly visible data and openness to dialogue (Retailer C), making order quantities easier to estimate. In contrast, limited communication, short order lead-time and high variation in demand that is difficult to predict because of fluctuating ordering patterns (Retailer F) potentially leads to disruption of the production schedule, thereby reducing efficiency and increasing waste.

SID 5 (Rev. 3/06) of 28

Table 5.1 Comparison of Case B’s Forecast Error/Accuracy on Retailers’ Orders During a 17-Week Period

Retailer

No of

SKUs

Average Order Lead-Time

(hours)

Forecast Error/Accuracy

Mean Percent

Error (MPE)

of Forecas

t

Rank

Mean Absolut

e Percent

Error (MAPE)

of Forecas

t

Rank

Mean Absolut

e Percent Variatio

n (MAPV)

in Deman

d

Rank

Percent Variation

in Demand Explained (PVE)

Rank

A < 10 < 12 13.3 4 18.3 5 25.0 1 26.7 6

B < 10 < 24 5.6 3 12.1 2 55.8 3 78.3 3

C < 20 < 12 4.1 2 12.2 3 87.9 5 86.1 2

D < 20 < 12 3.8 1 11.7 1 89.1 6 86.9 1

E < 20 < 24 13.8 5 20.3 6 55.5 2 63.4 5

F > 20 < 12 14.0 6 17.4 4 76.8 4 77.3 4


6. Survey of Chilled Food Manufacturers

6.1 Questionnaires and DistributionIn order to gauge whether the waste issues highlighted by this research were widespread in the chilled food sector generally, a survey of chilled food manufacturers was carried out. A questionnaire (Appendix 2) was sent to General Managers at 50 sites. Respondents could not be identified from the completed questionnaires. Twelve responses were received from the following chilled food sectors:

Five sites producing ready meals and/or convenience foods; Three sites producing fruit and/or vegetable products; One site producing prepared salads; Two sites producing sandwiches; One site producing meat products.

6.2 Survey ResultsProduction levels ranged from 130 to 500 tonnes per week, averaging about 320 tonnes. Weekly food waste ranged from 6 to 135 tonnes with a median value of 17.5 tonnes. Food waste as a percentage of production ranged from 2% to 33%, with a median value of 6%.

Surveyed sites were asked to provide estimates of what proportion of their food waste was due to specific causes (see Appendix 2). Waste from production related causes (trimmings/peelings, machine waste and production defects was highly variable, ranging from 10 to 89% of total food waste, with a median value of 55% (Figure 6.1). Production defects ranged from 1 to 40%, with a median value of 7.5%. Machine waste was similar with a range of 1 to 35% and a median value of 8%. Both these types of waste tended to be higher at sites producing more complex products, such as ready meals and sandwiches. When asked how often late changes by retailers to final orders disrupted production schedules, six said daily, four said a few times per week, one said several times per month, and one said only occasionally.

Waste from out of life (OOL) items also varied widely between sites, ranging from zero to 70% of total food waste with a median value of 27% (Figure 6.2). The median value for OOL raw materials was 11%, whereas the medians for OOL intermediate and finished products were both 5%. A small proportion of OOL waste is probably due to poor stock rotation but the remainder is likely to have arisen from over-supply or over-production and therefore could be seen as an indicator of how successful the sites were at forecasting customers’ orders, with 11

SID 5 (Rev. 3/06) of 28

of the 12 sites apparently having difficulties to varying degrees. All 12 sites generated their own forecasts. Only one site did not receive customer-generated forecasts but had access to their customers’ retail sales data. Of the remaining eleven, seven sites received forecasts from only some of their customers, with one site receiving no retail sales data; the other four sites received forecasts from all customers, all of whom shared their retail sales data.

Figure 6.1 Percentage of food waste from production-related causes at the sites of survey respondents

0102030405060708090

100

Ready MealsReady MealsReady MealsReady MealsReady MealsSandwichesSandwichesSalads

Fruit & VegFruit & VegFriut & Veg

Meat

% of Food Waste

Trimmings/Peelings Machine Waste Production Defects


Figure 6.2 Percentage of food waste from out of life (OOL) items at the sites of survey respondents

0102030405060708090

100

Ready MealsReady MealsReady MealsReady MealsReady MealsSandwichesSandwichesSalads

Fruit & VegFruit & VegFriut & Veg

Meat

% of Food Waste

OOL Raw Materials OOL Intermediate Products OOL Finished Products


When asked about forecast accuracy, the one site that did not receive customer forecasts, a ready meals site, admitted that their own forecasts were not at all accurate, and here OOL items comprised 70% of food waste, being mostly finished products. Another site (fruit and/or vegetables) said that neither their own forecast nor those of their customers were at all accurate, although OOL items comprised only 23% of food waste whereas 70% was from trimmings and peelings. Of the remaining ten sites, three said their own forecasts were extremely accurate and their customers’ forecasts were quite accurate but, even so, at two of these sites (both ready meals) OOL wastes stood at 50% (7.5 tonnes) and 55% (27.5 tonnes), with the highest proportion in both cases from intermediate products. Seven sites said both their own forecasts and those of their customers were quite accurate. Of these, six said they incurred OOL wastes, with totals ranging from 10 to 70% and a median value of

SID 5 (Rev. 3/06) of 28

26%. Tonnages ranged from 0.9 to 13.5 tonnes, with a median value of 5 tonnes. With such high proportions of OOL wastes at some of these sites, it is difficult to see why they consider their forecasts as quite accurate. Are they admitting extremely poor stock rotation? Are the actual tonnages at some sites so low that they cause little concern, even though the percentage is high? Or do they not equate over-supply of materials and over-production of intermediate and finished products with forecasting error?

The questionnaire included nine statements on potential options for reducing waste, and respondents were presented with a Likert scale to indicate their strength of agreement or disagreement (see Appendix 2). There was widespread agreement (83%) that fewer late changes to orders and fewer production changeovers would reduce waste. It was found during Case Study B that these two causes of waste are often linked. There was the same level of agreement that extended shelf lives would be helpful. There was almost as much agreement (75%) that better supply chain coordination and more involvement in generating final order quantities were required. Slightly fewer (66%) agreed that more information on retail sales would be helpful. The quality of raw materials caused concern to 58%. Only 50% agreed that fewer SKUs would reduce waste, perhaps surprisingly because the number of these produced often influences the number of changeovers. However, as found during Case Study B, changeover waste is affected by levels of setting-up waste and by the number of wash-outs and clean-downs required between runs, which depends upon the extent to which products can ‘follow on’.

The respondents were invited to make additional comments but only five chose to do so.

Type of supplier Comments (exact wording)Meat products “Weather forecasting – this is the biggest single factor affecting our sales

demand”.

Fruit/vegetables “More flexibility with recipes”.

Ready meals “- Reduce the frequency of redeveloping products - More flexibility from supply base, (i.e.) Min order quantity, lead times…(but this comes at a cost)”.

Ready meals “Production cycle is 3 days (placing order for raw materials, components etc > cooking > chilling > packing > despatch) where as receipt of order to despatch is 6 hours – hence waste. Individual opinion – not the company”.

Dressed salads “Total life of product & customer stock holding are the key to reducing waste. Because any +/- sales v. forecast takes four days within the supply chain to reach the customer it is imperative product life is increased to avoid the knee jerk reaction we constantly see. This would also mean stores increasing their potential stock holding space”.

The comments on minimum order quantities, lead-times and shelf life re-affirm some of the issues found during this research. The suggestion about increased customer stockholdings may just move OOL wastes from one location (the manufacturer) to another (the retail outlet), but it might provide retailers with an incentive to review their ordering and shelf replenishment procedures.

To summarize, the results of the survey echoed many of the opinions of senior managers at Case A and Case B regarding the quality of materials, visibility of data, late changes to orders, and collaboration and coordination throughout the supply chain.

7. Summary of Key Findings

This section reviews the key findings, from the retailers’ perspective, as gauged indirectly from the British Retail Consortium (BRC) and from conference presentations by supermarket representatives, and from the suppliers’ perspective. Then the environmental implications of the case study findings are scaled to UK level.

7.1 The Retailers’ PerspectiveFrom a meeting with representatives of the BRC, the key findings on waste at the retail outlet were as follows.

Waste is driven by:- Food safety concerns;- Legislative requirements;- Marketing;

SID 5 (Rev. 3/06) of 28

- Poor vendor compliance;- Consumer expectations of constant availability and uniformity of appearance, which is often mistaken for

higher quality. The major issue for retailers is to ensure availability to prevent a loss of customer loyalty. This can cause

over-supply and waste.

This last point was reinforced to some extent by an analysis of substitution by consumers of certain pastry products with similar products, showing there was a high level of non-substitution (Norton, 2007).

The views expressed in conference presentations by senior managers from the two largest UK supermarket chains provided insights into their relationships with suppliers.

Collaborative partnerships are helpful for improving demand management and reducing inventories but are largely reserved for the major branded suppliers, not the SMEs typical of chilled food manufacture.

Suppliers are expected to turn the data provided on I.T. systems into “useful information”. This can prove difficult for the small companies with low profit margins and limited resources that are typically own-label suppliers.

The responsibility for “getting it right” is pushed back onto the suppliers, who must bear the risks from holding buffer stocks to ensure availability with the possibility of non-requirement and, in the case of chilled foods, the costs attributable to refrigerated storage.

7.2 The Suppliers’ PerspectiveThe overall situation faced by manufacturers and suppliers of chilled food products to the major retailers is as follows:

Continuous improvement efforts tend to be thwarted by the need to meet daily orders. Order lead-times are generally less than 24 hours but production (or supply) lead-times can be several days,

so production has to start in advance of receiving final orders, with quantities based on forecasts that are inevitably inaccurate.

Under-estimates can cause difficulties in planning production schedules for high numbers of SKUs because additional short production runs must be slotted in to meet shortfalls and shorter runs cause more changeover waste in terms of both physical wastes and production time.

To ensure availability in response to retailers’ volatile ordering patterns and to reduce production-related waste from short runs, suppliers are obliged to hold buffer stocks in spite of the associated risks and costs due to being short shelf life products.

Key factors affecting waste are:- The type of product (whether cooked or uncooked);- Length of shelf life (cooked products have longer shelf lives);- Length of order lead-time versus production lead-time (fresh produce and cooked products tend to

have longer supply/production lead-times than uncooked products made from bought in, ready prepared ingredients).

7.3 Case Study FindingsEach of the case studies raised specific issues as follows.

Case Study 1: Cherry Tomatoes The quality of materials was a major issue, especially in relation to imports from distances further than

southern Europe, probably due to greater deterioration during longer journeys. The order lead-time of less than 24 hours was greatly exceeded by the lead-time for supply of at least 7 days

to reach the retail outlet. Buffer stocks were held for 3 days on average to ensure temperature compliance and, more specifically,

availability so as to compensate for forecast inaccuracy and order volatility. Overall waste of cherry tomatoes at the Pack House was 5.5% of imports, but might have been higher if

surplus stocks had not been sold to wholesale markets. Estimated waste of cherry tomatoes at the retail outlet was 4.5% of imports and was probably the result of

over-supply to ensure on-shelf availability, although wastage per SKU per branch per week would have amounted to only a few cases.

Case Study 2: Pastry Products The high levels of production waste were predominantly due to the nature of the product. Since carrying out

the case study, modifications to equipment have taken place, which have reduced this specifically product-related waste.

Buffer stocks were held at all stages of production:- To facilitate scheduling of a large number of SKUs across a limited amount of production equipment;- To reduce production-related waste by scheduling longer runs with fewer changeovers;

SID 5 (Rev. 3/06) of 28

- To ensure availability so that unexpectedly large orders could be met. Total food waste was equivalent to 24% of production but this falls to 11.5% when purely equipment-related

pastry trimmings are excluded. Rejected items were the second highest cause of waste, equivalent to 6% of production. Out of life wastes were normally very low but were significantly increased in specific instances by the

breakdown in a retailer’s usual best practice procedures. Waste was exacerbated by:

- Order lead-times of less than 12 hours from most retailers compared with production lead-times of up to 3 days;

- Volatile order quantities that increased forecast error;- Large discrepancies between retailers’ estimates and actual orders;- Insufficient data on retailers’ I.T. systems;- The unwillingness of some retailers to enter into a constructive dialogue;- Lack of communication on weekends and Bank Holidays.

Senior managers held the following views:- The more data made accessible by retailers, the more accurate the forecast will be;- Order lead-times of Day 1 for Day 3 or later would help in planning efficient production runs, reducing

stock rolled forward and providing the end-consumer with a fresher product;- Extending shelf-life would not in itself be helpful unless carried out in conjunction with extended order

lead-times and more tightly controlled shelf replenishment at stores;- Promotions encourage waste but can be important sale opportunities for suppliers.

Case Study 3: Automatic Vending of Chilled FoodsIn many ways this case study represents a microcosm of the waste issues surrounding chilled foods: the provision of short shelf life items in larger than demanded quantities, in order to make production commercially viable and to ensure availability of a wide choice of items at point of sale, coupled with poorly controlled stock rotation inevitably leads to waste. Forecasting, although never totally accurate, can greatly reduce waste if not impeded by the behaviour of various actors in the supply chain.

7.4 Survey FindingsThe survey results from twelve chilled food manufacturers reinforced the findings of the case studies.

Food waste, as a percentage of production, ranged from 2% to 33% with a median value of 6%. Machine waste (from set-up, changeover, cleaning, etc.) ranged from 1% to 35% of total food waste, median

8%. Production defects (from rejects, damage, errors, etc.) ranged from 1% to 40% of total food waste, median

7.5%. Total out of life (OOL) wastes ranged from zero to 70% of total food waste, median 27%. The median value

for OOL raw materials was 11%. The median values for OOL intermediate and finished products were both 5%.

Some proportion of machine waste will be caused by additional changeovers due to discrepancies between forecasts and actual orders but, as found from the case study on pastry products, obtaining a value for this would require detailed monitoring of various items of data over an extended period. Out of life intermediate and finished products will have incurred significant energy use from cooking processes (if required) and refrigerated storage.

The majority of survey respondents considered their forecasts to be either quite accurate or extremely accurate, and those of their customers to be quite accurate, in spite of the occurrence of OOL wastes, which in some cases were high in terms of tonnage if not percentage. These particular companies did not appear to acknowledge that forecast error, for whatever reason, could be a significant driver of OOL wastes. The quality of raw materials caused concern to over half the respondents. Only 50% agreed that fewer SKUs would reduce waste. However, there was widespread agreement that waste could be reduced by fewer late changes to orders, fewer product changeovers during production, better supply chain coordination, greater involvement of suppliers in generating final order quantities, and more information on retail sales.

7.5 Wider Environmental ImplicationsThe findings on wastes and greenhouse gas emissions from the case studies were scaled to UK level.

For UK imports of Spanish tomatoes (all varieties) totalling 190,000 tonnes annually (Anon., 2007), a similar level of waste to that in the case study would cause about 14,700 tonnes of wasted tomatoes in the supply chain before reaching the end-consumer. About 12,200 tonnes of cardboard would be required to import the total quantity, of which around 940 tonnes would have arisen from the quantity wasted. Approximately 7,000 tonnes of CO2 equivalents per year could be attributed to the life cycle of the wasted tomatoes (if landfilled) and to the cardboard packaging arising from those wastes (if recycled).

SID 5 (Rev. 3/06) of 28

For UK sales of all brands of pastry products in the same varieties as those analysed in the case study, amounting to around 12,900 tonnes annually, and assuming similar levels of waste to the case study, manufacture of these products would cause approximately 2,900 tonnes of food waste, 40 tonnes of plastic waste, 35 tonnes of cardboard and paper waste, and 9 tonnes of metal wastes, and would incur 70 million litres of water consumption and 5,400 tonnes of CO2 emissions from energy use per year. The quantity of food waste would be closer to 1,500 tonnes per year if the proportion calculated as being due to product-related pastry trimmings in the case study were excluded. These figures do not include wastes and emissions at the other life cycle stages of raw material production, primary processing, use, disposal, or the intervening transport stages.

A reduction throughout the chilled food sector in the requirement for chilled storage areas by holding lower levels of buffer stocks would save, at the very least, tens of thousands of tonnes of CO2 emissions annually from energy consumption but smaller buffer stocks would require greater certainty in order quantities.

7.6 Limitations of FindingsIt is important to point out certain gaps and limitations in the research findings. Perhaps most importantly, it was not possible to analyse an entire value stream. DEFRA had requested a post farm gate investigation, up to and including retail outlet, but this was not achieved. The post factory waste data obtained was minimal; in Case Study A, the quantity wasted after leaving their site had to be assumed from the difference between the quantity ordered, assumed delivered, and the quantity sold at the checkout, as indicated by EPOS data; in Case Study B, no EPOS data was provided for reasons of customer confidentiality, so only intra-company waste was analysed; and, lastly, it proved impossible to engage with retailers in order to quantify and attribute the waste arising at depots and stores, and during distribution. Stakeholders upstream were also omitted as it would have been difficult to ensure the anonymity of case study participants.

Another important limitation is the small sample size. Had the three abortive case studies also been completed (a supplier of processed meat products and two ready meals manufacturers), a broader picture of the chilled food sector would have been obtained, particularly because ready meals have the largest share of the retail market. Ready meals might also be more vulnerable to wastage than the pastry products analysed. The survey of chilled food manufacturers largely supported the findings from the case studies but, once again, the number of respondents (12) was small when compared to the total number of chilled food manufacturers of all sizes, estimated as hundreds by Goodburn (2007).

The case studies themselves may also have contextual limitations. Were these companies typical of other suppliers of similar products? Was there anything particular about the mode of operation that made either company non-representative? Certainly the quality drive by Case A’s customer had increased their waste levels substantially, although for the particular product analysed, cherry tomatoes, the exceptionally high levels of waste were driven by their increasing popularity, which forced Case A to source from further afield and incur higher wastage levels when Spanish supplies were insufficient. The high level of wasted pastry trimmings at Case B, although largely inevitable because of the nature of the product, could be reduced by the purchase of new equipment, which has occurred since the case study was carried out; the author has no information on the reduction in waste that this might have achieved.

In spite of these limitations, the qualitative evidence on the drivers of waste, particularly at the supplier-retailer interface, are indisputable even if the quantitative evidence on wastage levels is somewhat meagre. However, the additional data obtained from the survey of chilled food manufacturers showed that the quantities of waste arising due to various causes can be highly variable, even between manufacturers of similar types of product.

8. Potential Opportunities for Reducing Waste

Both physical and operational wastes, and the ensuing environmental impacts, could be reduced by the following improvements to current practices within the chilled food sector.

Retailers need to provide data that includes (as a minimum requirement) retail sales, stock levels and wastage levels.

The data provided has to be reliable in order to minimize under- or over-production occurring at the manufacturer.

There is a need for greater dialogue between retailers and suppliers, especially for smaller companies who currently lack the resources and expertise to turn data into “useful information”. Some minimum level of communication over weekends and Bank Holidays would also be helpful.

Ordering patterns need to reflect the underlying stability in the weekly sales of most food products, rather than volatile daily sales.

SID 5 (Rev. 3/06) of 28

Final order quantities need to be confirmed further in advance than at present so as to minimize disruption to efficient production schedules and thereby reduce production-related waste. This would also reduce the size of buffer stocks currently held as a safeguard against order uncertainty.

Producing to demand rather than to an inaccurate forecast still requires suppliers to improve their efficiency by cutting lead-times where possible, perhaps by reviewing the length of time spent in storage by intermediate and finished products, although this is also partially related to the number of SKUs produced in comparison to the number of production and packing lines available.

Throughout the chilled food supply chain, there needs to be a greater, combined awareness of Lean techniques, in order to improve efficiency, and the environmental implications of both current practices and future changes to those practices, in order to optimize environmental performance.

9. Constraints on Reducing Waste

The conceptual framework in Figure 2.2 showed the ideal sequence of events that brings about the changes in behaviour necessary for achieving the desired future state, where operational and environmental wastes are reduced throughout the supply chain in a cycle of continuous improvement. This research has highlighted numerous reasons why this ideal situation is difficult to attain in the chilled food sector. Undoubtedly there are drivers for change, in the form of legislation on environment, human health and food safety, and also incentives, for example, reduced costs. A further incentive is improved efficiency that raises service levels, ensuring satisfied customers and continuing orders. The adoption of Lean principles can help to improve efficiency but this requires training so that the concept is fully understood by all levels of personnel, from senior management down to the shop floor, engendering a culture of empowerment and team-working in order to solve problems. In the wider supply chain, collaboration and coordination, for example, by the sharing of information and the use of CPFR, are essential for the improvements in demand management that would prevent over-production. The outline of conventional VSM in Section 2 also highlighted the fact the whole supply chain needs to focus on one single definition of value, that is, value as specified by the final consumer. However, the environmental performance of the entire supply chain must also be addressed to ensure that a Lean implementation does not worsen environmental impacts and fail to meet sustainability criteria. (References?)These drivers, incentives and facilitators should reinforce the resolve to plan and implement the necessary process changes identified as a result of the raised awareness brought about by mapping and measuring performance indicators in the current state and prompt the changes in behaviour required for the realization of the future state where operational and environmental wastes are reduced. However, the observational site visits to chilled food manufacturers and the case studies carried out during this research, in addition to other anecdotal and empirical evidence, indicate that there are a number of internal and external factors that constrain implementation plans and prevent the necessary changes in behaviour.

The internal constraints include the following:

Time- Working in ‘fire-fighting mode’ as a result of short order lead-times tends to thwart improvement efforts.

Practicability- Are the process changes identified feasible? For example, if buffer stocks were reduced, would the

current set-up of chill rooms allow any to be switched off so that energy consumption is reduced? Costs vs. benefits

- Would the costs of process changes exceed the financial benefits derived? This is especially pertinent in an industry characterized by low profit margins.

Expertise- Staff must acquire skills in Lean principles, value stream mapping and environmental awareness.

Resources- Does the company have sufficient human and financial resources to identify and implement changes,

whilst continuing to meet the daily production schedule?

The external constraints are likely to be even more difficult to overcome and include the following:

Supply chain relationships- The major retailers are in a position of dominance; collaborative relationships are usually reserved for

large, branded suppliers whereas most chilled food manufacturers are SMEs; order lead-times are too short to allow production to demand; inaccurate estimates are provided and order quantities can be highly variable; insufficient data on demand and wastage is made available to suppliers; late changes to orders disrupt production schedules.

Consumer attitudes:

SID 5 (Rev. 3/06) of 28

- Especially those relationg to on-shelf availability, willingness to accept substitutes, and uniformity of appearance as a misleading indicator of quality.

The everyday realities faced by chilled food manufacturers that have been identified from the empirical evidence are reflected in the revised conceptual framework shown in Figure 8.1. The initial conceptual framework (Figure 2.2) showed arrows of equal thickness, denoting determination to implement the ideal future state, in which there is a collaborative effort to reduce operational and environmental wastes throughout the supply chain in a culture of continuous improvement. The revised conceptual framework has increasingly faint, fragmented arrows reflecting the real-world situation. The mapping and measuring of performance indicators raises awareness of the required process changes, and the drivers, incentives and facilitators ought to provide the impetus and capability to bring about the necessary changes in behaviour, but the planned changes are incomplete (fragmented arrow) because internal and external constraints limit the possible behavioural changes to the ‘easy fixes’ within the company, so that operational and environmental wastes are reduced to only a limited extent (fainter, more fragmented arrow) and continuous improvement falters (very faint, very fragmented arrow) because the external constraints impede the more difficult, but more effective, supply chain solution.

Figure 8.1 Revised Conceptual Framework

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.


9. Further Research

The findings from this research have the potential to contribute to the enhancement of sustainability within the chilled food sector in two ways:

a) By identifying failings at the interface between retailers and food manufacturers as a significant driver of waste;

b) By developing a diagnostic tool that is of practical value to industry.

However, further research is warranted in the following areas:

SID 5 (Rev. 3/06) of 28

Extension of the SVSM technique upstream into primary production and downstream into retailing. The problems associated with this should not be under-estimated, given the difficulty invariably experienced when trying to engage with retailers.

Improving the ease-of-use of life cycle assessment techniques for non-specialists so that it might be integrated with the SVSM approach to give a more comprehensive indication of the potential environmental impacts of proposed operational changes.

Finding more accurate estimates for energy consumption for specific points in the production and distribution processes, for example, the energy consumption associated with cold storage at manufacturing sites.

Exploring the potential for technological solutions to extend the shelf life of chilled products, and estimating the effect this might have on order lead-times and the quantity of waste generated, particularly at retail store level.

Breaking down the barriers that currently exist between different functional departments within and between businesses, which result in fundamental misunderstanding of the implications (economic and environmental, and upstream and downstream) of changes to discrete decisions. Investing in better relationships and improved communications could go a long way to reducing operational and environmental wastes caused by failings at the interface between buyers and sellers.

SID 5 (Rev. 3/06) of 28

References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.

Anon. (2002a) World Summit on Sustainable Development 2002: Contribution by the UK Food and Drink Manufacturing Industry, London, Food and Drink Federation.Anon. (2002b) Lifecycle Environmental Comparison: Virgin Paper and Recycled Paper-Based Systems (White Paper No. 3), New York, USA, Environmental Defense Fund, URL: http://www.environmentaldefense.org/documents/1618_WP3.pdfAnon. (2006a) The Lean and Environment Toolkit Version 1.0, US Environmental Protection Agency, URL:http://www.epa.gov/lean/toolkit/lean_environment_toolkit2.pdfAnon. (2006b) Chilled Food Manufacturing Waste Minimisation Study, Redcar, PICME/FPFP (for DEFRA)Anon. (2006c) International Demand Planning and Forecasting, Watford, Institute of Grocery DistributionAnon. (2007) Tomato Facts: Market Information, UK, British Tomato Growers’ Association, URL: http://www.britishtomatoes.co.uk/newsite/facts/info.htmlAnton, A., Montero, J., Munoz, P. and Castells, F. (2005) Identification of the Main Factors Affecting the Environmental Impact of Passive Greenhouses, Acta Horticulturae, 691, 489-494Brierley, P. and Spring, J. (2006) Lean Collaborative Planning, Demand Planning and Forecasting Conference, London, Institute of Grocery and Distribution.Chockalingam, M. (2003) Forecast Accuracy and Safety Stock Strategies, Lexington, USA, Demand Planning LLC, URL: http://www.demandplanning.net/documents/dmdaccuracywebVersions.pdfCrumrine, B. Decker, S., Loughman, E. and McMullan, R. (2005) Environmental Packaging Guideline for the Electronics Industry Version 2.0, Santa Barbara, USA, University of California, Donald Bren School of Environmental Science and Management, URL: http://www2.bren.ucsb.edu/~green-pkg/EPG.pdfDEFRA (2001) Environmental Reporting: General Guidelines, UK Crown copyright.Ellis, A. (2006) Managing Retailer Demand, Demand Planning and Forecasting Conference, London, Institute of Grocery and Distribution.Fisher, K. (2006) Impact of Energy from Waste and Recycling Policy on UK Greenhouse Gas Emissions: Final Report, London, Environmental Resources Management (for DEFRA).Gilliland, M. (2002) Is Forecasting a Waste of Time, Supply Chain Management Review, July/August 2002, 16-23.Goodburn, K. (2007) Personal Communication, Kettering, Chilled Food Association.Guinee, J (Ed.) (2002) Handbook on Life Cycle Assessment: Operational Guide to the ISO Standards, Dordrecht, Netherlands, Kluwer Academic Publishers.Jones, D and Womack, J. (2003) Seeing the Whole: Mapping the Extended Value Stream Version 1.1, Brookline, USA, Lean Enterprise Institute.Lapide, L. (1998) Forecasting is About Understanding Variations, Journal of Business Forecasting, Winter 1998-99, 29-30Lapin, L. (1980) Management Science for Business Decisions, New York, USA, Harcourt Brace Jovanovich Inc.Martinez-Inchausti, A. (2006) Personal Communication, London, British Retail Consortium.Mason, R., Simons, D., Peckham, C. and Wakeman, T. (2002) Wise Moves Modelling Report: Life cycle modelling CO2 emissions for lettuce, apples and cherries, UK Department of Transport.Norton, A. (2007) Sustainable Value Stream Mapping as Technique for Analysing and Reducing Waste in the UK Chilled Food Sector, PhD Thesis, University of London, Imperial College, Centre for Environmental Policy (Applied Economics and Business Management Research Section).Ohno, T. (1988) Toyota Production System: Beyond Large-Scale Production, Portland, USA, Productivity Press.Rother, M. and Harris, R. (2001) Creating Continuous Flow: An Action Guide for Managers, Engineers and Production Associates Version 1.0, Brookline, USA, Lean Enterprise Institute.Rother, M. and Shook, J. (2003) Learning To See: Value Stream Mapping To Create Value and Eliminate Muda Version 1.3, Brookline, USA, Lean Enterprise Institute.Swoffer, K. (2006) Personal Communication, London, British Retail Consortium.Taylor, D. and Fearne, A. (2006) Towards a Framework for Improvement in the Management of Demand in Agri-Food Supply Chains, Supply Chain Management: An International Journal, 11(5), 222-236.Womack, J. and Jones, D. (2003) Lean Thinking, London, Simon and Schuster UK Ltd.Yin, R. (1994) Case Study Research: Design and Methods Second Edition, Thousand Oaks, YSA, Sage Publications Inc.

SID 5 (Rev. 3/06) of 28

http://www2.bren.ucsb.edu/~green-pkg/EPG.pdf

http://www.demandplanning.net/documents/dmdaccuracywebVersions.pdf

http://www.britishtomatoes.co.uk/newsite/facts/info.html

http://www.epa.gov/lean/toolkit/lean_environment_toolkit2.pdf

http://www.environmentaldefense.org/documents/1618_WP3.pdf

SID5_ft0348 - Secretary of State for Environment, Food...

Documents

Transcript of SID5_ft0348 - Secretary of State for Environment, Food...