Practical implementation of production planning and control concepts in SMEs and MTOs: an introduction to the special issue

Context

A major branch of the available literature on production planning and control (PPC) concepts is theoretical and focused on developing more and more sophisticated methods of solving problems through simulation and modelling. In becoming more sophisticated, this work has moved further away from the type of simple, workable solution that can be implemented in practice. This can be seen, for example, in the literature on determining due dates and shop floor scheduling. Far less research is field-based and concerned with furthering the development of PPC concepts through implementation, particularly concepts of relevance to small and medium sized enterprises (SMEs) and manufacturers of customised products, such as make-to-order (MTO) companies. It is this that the special issue focuses upon.

The characteristics of SMEs and MTOs present particular challenges to the implementation of PPC concepts. SMEs often have limited financial resources and IT infrastructure and are reliant on a small number of key individuals who cannot be relieved of their daily roles and responsibilities to focus on a project. Meanwhile, MTOs produce a wide variety of products in low volumes (sometimes one-of-a-kind products), making lead times difficult to predict and shop floor routings highly variable. The characteristics of SMEs and MTOs limit the number of suitable concepts and make successful implementation a major challenge. This is heightened by differences between simulation and practice, such as the following:

• Model assumptions: Assumptions used in simulations do not necessarily translate to the real world. For example, the production processes found in practice tend to be much more sophisticated than those tested in simulation. Production may take place on parallel machines, setups may take place offline, and product structures may be more complex. Some PPC concepts require more data than is routinely available in many SMEs and MTOs. For example, for a one-of-a-kind job, it may be unrealistic to expect detailed data on processing or setup times to be available in order to calculate a due date when the job has not been produced before. At the shop floor level, some methods require quick and accurate feedback information on the progress of jobs, but this can be difficult to provide in SMEs with limited resources.

• Underestimating the human dimension of production systems: The human element of real-life production systems impacts a wide range of issues, including the capacity of the shop and the flow of orders. In simulation, capacity can arguably be much more flexible in the short term than in reality. Moreover, unlike many models, in reality operators do not work at a uniform rate or strictly adhere to sequencing rules; they might slow down towards the end of the week to protect overtime, process the order for the customer that ‘shouts the loudest’, or work on the order which will get them the biggest bonus (rather than the order with the most urgent due date).

• Unfamiliarity with the PPC concept: If the PPC concept is complex or unfamiliar to practitioners, providing training is important; this is not required in simulation and is sometimes underestimated in practice. If adequate training is not provided, the user of the PPC concept is likely to make ill-informed decisions; this creates decision ‘chaos’ that is not a factor in simulations.

• Organisational ‘unreadiness’ for implementation: A good fit between the characteristics of an organisation, e.g. in terms of shop configuration and volume/variety, and a PPC concept is important, but there are also many other issues to consider which are harder to judge. These include company culture, political landscape, project commitment and willingness to change. These issues do not have to be considered in simulation, but can have a major bearing on the outcome of an implementation project.

This special issue focuses on new contributions to the development, implementation and popularisation of PPC methods in practice through field research, with a particular focus on SMEs and MTOs. It is motivated by a need: to engage with practice in order to develop methods that not only produce good results, but which solve problems in a way that can be implemented in practice; to build up a greater body of empirical evidence and knowledge on the implementation of PPC methods of relevance to SMEs and MTOs; and, a desire to contribute towards addressing questions such as the following:

• How can PPC theory be aligned with the characteristics of the SMEs and MTOs that researchers encounter in practice?

• How can a compromise be reached between the performance benefits of sophisticated PPC theory tested in simulation and the simplicity of the PPC methods and decision-making processes often preferred in practice?

• What are the barriers to, and the enablers of, the implementation of PPC concepts in SMEs and MTOs?

• Which issues encountered when implementing concepts in SMEs and MTOs are broadly relevant to all PPC concepts and which are concept-specific?

• What insights can be provided to help facilitate more successful PPC implementations in SMEs and MTOs in practice?

Organisation of the special issue

Nine papers are contained within this special issue; some focus on SMEs, some on MTOs and some on both. Some explore and adapt the use of methods designed for large companies to SMEs; others, the use of methods for make-to-stock production to MTO production. The papers cover a range of new and established approaches to PPC that vary in sophistication, including paired cell overlapping loops of cards with authorisation (POLCA), workload control (WLC), constant work-in-process, and advanced planning and scheduling (APS) systems. They also encompass a range of research methods differing in the level of participation and observation, from cross-sectional to longitudinal work and even to the use of secondary empirical data. All of the papers are geared towards improving the application of PPC concepts to the company characteristics, production planning environments and problems found in practice.

The first paper, by Land and Gaalman, presents case study evidence on the progress of orders in MTO SMEs. Data is collected over a minimum of 3 months using a scanning system implemented in seven companies. Hence, the paper is a relatively rare example of longitudinal case study research that involves multiple companies. The authors analyse logistic performance using throughput diagrams to diagnose problems related to average lateness and order progress diagrams to diagnose problems related to the variance of lateness. By examining the root causes of logistic performance problems and identifying cross-case commonalities, the authors identify a number of implications for the future development of PPC concepts in order to better meet the needs of MTO SMEs.

The authors find that most logistic performance problems can be anticipated before an order has been released onto the shop floor. They also highlight the importance of providing capacity planning overviews to support sales decisions at the customer enquiry stage and stress the importance of controlling the pre-production phase in MTO companies for accepted orders; pre-production includes design and engineering, process planning, and the procurement of customer-specific components. If this stage is not controlled, it can undermine the excellent work previously conducted by PPC researchers at the later planning and control stages, such as of order release and shop floor dispatching. The authors have a particular interest in the WLC concept, but the work has clear implications for the development of all PPC concepts for MTO companies.

The second paper, by Perona, Saccani and Zanoni, attempts to develop a solution that bridges the gap between inventory planning theory and the policies typically adopted in practice. The authors note that researchers have developed methods based on optimisation that are often too complex for practical application or which do not consider all of the factors that are relevant to inventory planning in practice. It is argued that the inventory planning methods applied in practice are often flexible and ad hoc; for example, whether a particular product is made to-stock or to-order may be based on common sense and experience, while lot sizes for stock items may be based on available warehouse space rather than inventory versus set-up costs. But in being ad hoc, decisions are often inconsistent.

The work is not only relevant to readers interested in inventory planning, it is also broadly interesting from a methodological point of view. Instead of beginning with theory and refining it for practical application, the authors' study is grounded in practice. They begin with an SME and work backwards, exploring how established inventory planning policies can be selected, combined and assigned to product groups according to their requirements. In doing so, the authors develop a solution that strikes a balance between ease of use and performance, building spreadsheets with simple data requirements rather than complex tools that are difficult to understand and maintain and which have interfaces that are unfamiliar to the user.

The third paper, by Wiers, presents evidence from two action research implementations of APS systems. The author explores how the implementation of an APS system places responsibility for planning and scheduling activities on a central scheduler. This can require significant and careful change management if responsibility was previously held on the shop floor, making the level of autonomy a key implementation issue for APS systems. In one of the cases the implementation is successful, partly because an understanding already existed on the shop floor of the need for a central, high level view for planning and control. As a result, implementation did not involve a radical change to the culture or PPC hierarchy within the company. In the other case the implementation was less successful, partly because operators were accustomed to having decision freedom on the shop floor. The schedules produced by the APS system at a central level where undermined on the shop floor at a local level because aligning this central computerised planning system with the localised shop floor control involved too big a cultural shift.

Local shop floor control can help to cope with change and uncertainty and motivate shop floor operators, while central control can lead to globally optimal, rather than local suboptimal, schedules. But ultimately, an agreement must be reached over what is scheduled by the APS system and what is performed on the shop floor. The paper provides a rare empirical study of a key issue affecting the implementation of APS systems and stresses the importance of the fit between the system and the planning hierarchy of a company. The case in which the implementation has been most successful is a flow shop, while the case in which implementation was less successful is a job shop. It is perhaps unwise to generalise based on two cases, but in addition to providing an insight into behavioural issues affecting APS implementation, the paper may also say something about the applicability of these systems to MTO companies, which often have a job shop configuration.

Lean principles have been most commonly adopted in low-variety/high-volume production environments. But through case study research, the fourth paper, by Slomp, Bokhorst and Germs, demonstrates that some of these principles can be implemented, or adapted for implementation, in a high-variety/low-volume MTO environment. After successfully implementing their ‘lean production control system’ in a small production unit, production flow times were reduced and stabilised. Overall lead times were not reduced significantly as the production time savings were largely shifted to the unit's order pool prior to release. However, holding orders in the pool for longer can provide flexibility as last minute changes to order requirements can be made at less inconvenience. By also lowering the level of WIP, an attempt can be made to ensure that shop floor operators are working on the most urgent orders; as a result, on-time delivery performance improved substantially. Notable factors which contributed to their success include the use of simple software tools and a focus on improving understanding and acceptance within the organisation through gaming.

The main focus of the work is on shop floor control; hence, the method is likely to require additional control at a higher level for MTO implementation. In the case described, although the initial production unit which Slomp et al. investigated is small, it is the part of a larger organisation with the features of a larger organisation, including access to an MRP system for higher level planning. Use of the lean production control system has begun to be extended to other parts of the organisation with similarly positive results.

The fifth paper, by Krishnamurthy and Suri, describes POLCA, a hybrid push-pull approach for high-variety or customised production environments, such as the MTO industry. The card-based shop floor control system is proposed as an alternative to pull systems such as the Kanban, which the authors explain was originally designed for repeat production with stable demand. It is claimed that the hybrid approach, where cards are assigned to pairs of cells to control the movement of materials between the two cells, combines the best features of pull and push systems. The authors outline a number of prerequisites for the implementation of POLCA. These include the need for: high-level MRP planning to support the use of POLCA for low-level shop floor control; manufacturing to be organised into cells; estimates of capacity and lead times to be available; and, a simple scheduling system that would incorporate release authorisation and determine production schedules. This shows similarities with the lean production control system described by Slomp et al., which includes a card-based pull system and is implemented in a high-variety environment with MRP at a higher planning level. MRP as a prerequisite to the implementation of POLCA can add an extra layer of complexity to implementation if it is not already in place and may limit adoption in SMEs. Similarly, Wiers explains that APS is often implemented by companies that use ERP systems.

The paper helps to further the implementation of POLCA in practice by outlining the steps that researchers and practitioners must take to embed the approach within an organisation. A four-stage procedure for implementing POLCA is outlined. This consists of an initial POLCA assessment to determine if the prerequisites for POLCA are present or if these must first be realised; a design stage in which, for example, the POLCA loops and number of cards required are determined, which can soon get complex for job shops with many routing permutations; an implementation stage in which, for example, a project champion is determined and training is provided; and, finally, a post-implementation stage where the outcomes of the project are evaluated and extensions are considered. The authors reflect on three successful implementations of POLCA. Outcomes include reduced lead times, increased on-time delivery performance and improved employee satisfaction. As with some of the other methods described in this special issue, implementation appears to often begin in one area of the factory before consideration is given to extending it to others; hence, this is another method that can benefit from gradual implementation.

Just-in-time (JIT) was once a radical approach generally confined to large manufacturing organisations. It is now considered more orthodox and applications of lean thinking have been widely reported, including in service contexts; however, its adoption amongst SMEs has received only limited attention. The sixth paper, by Dowlatshahi and Taham, investigates the barriers to, and enablers of, the implementation of JIT in SMEs. The authors acknowledge that while JIT may work well for many large companies downstream in the supply chain, it is likely to lead to inventory being held ‘somewhere else’ in the chain. For SMEs to deliver on a JIT basis and at short notice, it can be much more challenging for them to produce on a JIT basis too.

The paper is notable for its use of secondary empirical data, which varies in terms of the unit of analysis, from the case of a small manufacturer to a large-scale industry survey. But the contribution of the paper is not so much in the ‘cases' that the authors describe, but in the conceptual framework and hypotheses they develop. Barriers include issues around demand fluctuation and a lack of adequate training and investment. Enablers include empowering employees and overcoming resistance to change. Some of the issues are particularly relevant to the application of JIT, but others are of broader relevance. The authors argue that the work could form the basis of a diagnosis tool to help SMEs determine if, and in what ways, JIT is feasible for them. When combined with Slomp et al., the two studies provide an important insight into how lean principles and methods can be applied to both SMEs and MTOs.

The seventh paper is my own with Huang and Hendry; this describes the development of an end-user training tool embedded within a Decision Support System (DSS) that forms part of an implementation strategy being formulated for the WLC concept. Although we argue that it is a leading solution to the PPC problems of MTO SMEs, WLC is not as familiar to practitioners as some of the other concepts detailed in this special issue. The tool has been applied as part of an implementation project with a small subcontract MTO company. It was anticipated that the tool would play a key role in promoting acceptance and understanding of WLC in the case study company in order to improve the likelihood that decisions would be made in accordance with the concept, and it did. However, many other benefits were realised that we had not anticipated; for example, it played a part in selecting roles and allocating tasks to different users; it provided a focus for analysing and improving the fit between the system and the company's business processes; it informed the initial WLC parameter setting process; and, it will continue to have a part to play in the post-implementation stage by allowing users to simulate decisions off-line before making them in the ‘live’ system.

The paper contributes to the available literature on WLC by addressing some of the outstanding research questions raised by Hendry et al. (2008), including: ‘how can planners and other personnel be trained such that decisions will be taken in correspondence with the integrated WLC approach?’. As in Slomp et al., an interactive hands-on approach is adopted to training users. But in contrast to Slomp et al., and the paper presented by Perona et al., the software system does not take the form of a spreadsheet or other recognisable tool; it is a sophisticated piece of bespoke software with multiple screens and unfamiliar interfaces. This makes training and the promotion of user acceptance even more important. Simplicity is important where possible, but the key is ensuring that decision support tools and training are proportionate to the complexity of the PPC concept and the degree of organisational change required.

The penultimate paper, by Silva, presents a rich insight into the design and development of a planning and scheduling tool for a synthetic fibre production company. The tool strikes a balance between automating the scheduling task and leaving the human planner to produce the schedule manually. The solution claims to combine the strengths of theoretical models with the strengths of the ad hoc decision-making procedures typically found in practice. Instead of governing decisions, it seeks to provide decision support, allowing the human scheduler to intervene. To achieve this, the algorithm employed features hard and soft constraints, i.e. constraints that must be met and those which can be broken. The algorithm is able to quickly assess complex scheduling problems, while the human scheduler can intervene based on their knowledge of broader and less tangible considerations which require judgement. The human scheduler can consider factors that would be beyond the algorithm, such as their knowledge of the production process, e.g. whether to extend the lifetime of tooling to increase output at the risk of affecting quality, and wider issues, such as whether to delay particular orders to improve the overall plan or the need to satisfy the requirements of the organisation's sales department. The approach has been successfully implemented in the case study company. The production plans are quickly produced and are ultimately better than before–the human scheduler is able to dedicate less time to producing the plans and so has more time to revise and improve them. This could have implications for the future development of other PPC tools for complex production environments.

Most studies in this special issue are moving away from optimisation techniques. The final paper, by Verlinden, Cattrysse, Crauwels and Van Oudheusden, is an exception that is in sharp contrast to the approach adopted elsewhere in this collection of papers. The authors present the development of an optimal production planning methodology for sheet metal working SMEs. Based on a survey of over 100 sheet metal working firms, of which 70% are categorised as either MTO or engineer-to-order firms, the authors find that in practice: firms tend to base production planning on experience rather than complex algorithms; the experience of the human planner affects the plan developed; and, plans are often far from optimal. But in this complex environment, where efficient use of both metal and machine time is so important, the survey finds that an automated optimal solution would actually be welcomed by practitioners.

The authors' review of the literature indicates that there are models that can provide optimal plans for individual production stages, but they argue that planning decisions made at one production stage affect those made at another. Therefore, they extend optimisation to two stages, synchronising production plans across laser cutting and air bending. Then, in collaboration with sheet metal working SMEs, the authors incorporate practical considerations into their integer program, such as rush orders and the need to limit computational times for the solution to be useable in practice. The value of the solution is tested using data collected from SMEs, demonstrating its speed and superior performance over existing methods found in the companies.

Discussion

This introduction to the special issue has provided a detailed summary of the papers it contains. While the few cases on which each individual study is typically based limits the generality of the findings, collectively the special issue contributes to building up a larger body of evidence from which greater generalisation may be possible. Some of the recurring themes include: an ad hoc approach to decision making is typically found in SMEs; PPC implementation often relies on software tools which add to the complexity of implementation; there is a preference for simple, flexible, transparent, fast, and effective tools to support decision making in practice; data availability is limited in practice; and, adopting a gradual approach to implementation can be an important part of managing change.

Two broad approaches are adopted by researchers in the field studies described. First, there are the researchers that begin with an interest in a particular PPC theory and attempt to embed it within an organisation that they ‘choose’. And secondly, there are the researchers that begin with a particular company and select or adapt theory from a ‘toolbox’ of PPC principles as appropriate. Papers which adopt this second approach have to walk the tight rope between research and consultancy, but it is perhaps no coincidence that all of the studies which took this approach report successful implementations. Moreover, starting with a problem does not mean that a new contribution to knowledge cannot be made; it can lead to the development of more practice-oriented theory. Whichever approach is adopted, in general a compromise is sought, which moves theory closer to existing practices when practices are ‘good', stubborn or common, and moves current practices towards theory when current practices are poor, changeable or undermine a core element of the method.

The papers also highlight the complexities of PPC in SMEs and MTOs. But on many occasions, the complexities of PPC can be overcome; it is often the more cultural, behavioural and political issues within organisations that cause the biggest implementation problems. This is one reason why field research is risky–there were significant timescales involved in all of the field studies and no guarantee of positive results. In fact, drafts of several promising papers were withdrawn by prospective authors because: of implementation delays caused by lost project momentum, a change of management structure or data confidentiality issues; or, the collaboration with the company collapsed. But although field research can be ‘messy’ and high risk, it is worth the investment, and if a rigorous approach is adopted, rich and rewarding insights can emerge whether the outcome is positive or otherwise.

Future field research should be conducted which pursues the directions identified in the individual papers in this special issue, including the incorporation of pre-production stages within planning concepts to improve their applicability to MTO companies. Other suggestions include:

• Organisational fit: Research should be conducted which seeks to provide greater guidance to support the selection of an appropriate PPC method or combination of methods for a given set of company characteristics. Or, in reverse, provides greater support for how to choose a suitable company for a particular concept and research project. Further research is also required to provide more detailed strategies on how the methods described in this special issue can be embedded within organisations.

• Behavioural dynamics: Further research should be conducted which focuses on the behavioural issues which affect implementation, including how users interact with the computer systems that often support the use of PPC concepts in practice.

• Knowledge management: Given the limited data availability in some SMEs, further research should be conducted into how information and knowledge management in SMEs can be improved.

• Sustainability: Future research should conduct follow-up studies in which companies involved in previous implementations are revisited. This could consider, for example: if the implementation has been maintained or extended; if use of the method has evolved to better fit practice; if the method has been neglected over time as hype surrounding the project reduced or it became too time consuming to maintain; and so on.

• More realistic simulations: Insights from field research should be fed-back into simulation and modelling research–which also has an important role to play in the development of the field–in order to develop more realistic simulation experiments.

• Supply chain effects: SMEs play an important role in supply chains and are often positioned towards the upstream end of the supply chain, receiving rush orders from important and powerful customers with imminent non-negotiable due dates. Hence, issues beyond the boundaries of the focal firm, including the typical nature of relationships with customers, clearly affect the outcome of a PPC implementation project and should be given greater attention in future research.

• Service planning and control: Research is required which explores how PPC concepts can be adapted for application in more service-oriented contexts.

Final remarks

Call it what you will–‘where the rubber meets the road’ or ‘where the theory hits the fan’–but field work can provide an important scientific contribution both in testing and refining existing theory and in developing new theory. The large number of enquiries and submissions to this special issue demonstrated the level of interest that researchers continue to have for PPC. The final choice of papers reflects the breadth and depth of research of relevance to SMEs and MTOs that is emerging and the quality of the collaborative work taking place between industry and academia. I would like to thank all the reviewers and authors for their contributions to this special issue. I hope it will be of interest to all readers of the journal.