JIT Approach to Mass Customization: A Case Study

CARMO, Francisco Duarte de Castro Ferreira & GAVRONSKI, Iuri. JIT approach to mass customization: a case study. In: Production and operations management society: POM 2002 Annual Meeting. San Francisco: 2002.

Autores

Francisco Duarte C. F. Carmo
Iuri Gavronski

Mass customization is one of the main challenges to Operations Management, as it aims to deliver a wide variety of products keeping inventories and costs under control. Besides that, customers are always demanding smaller delivery times, which cannot be satisfied with final products inventories, due to the large number of different items that can be produced.

One approach to this problem could be a combination of product and plant design were the concepts of mass customization and postponement are combined with modular product design and the implementation of cellular manufacturing with Kanban production control, to allow a postponement strategy and JIT delivery of final products, with minimal inventories and productivity increase.

This case study was conducted at Vanbro Submersible Pumps Ltd., a submersible water pumps small manufacturer located in Brazil. The company has a wide variety of submersible pumps: almost 600, combining variables as wheel diameter (4” or 6”), required pressure and available electric power at the field. Besides that, customers are always demanding the smallest delivery times possible, sometimes less then 3 hours, as they run out of water when a pump breaks.

The following people from Vanbro made possible this case study: Edilson Klein, Euclides L. Brock, Fabiane N. Manfrão, Iara T. Martins, Ivo B. Vanzella, Jorge P. Rovadoschi, Marcos J. C. Kuhn, Paulo H. Kist, Paulo L. Gomes, Paulo R. P. Almeida and Valdinei Mendes.

Introduction

The early works on Manufacturing Strategy (Skinner, 1969, Hayes and Wheelwright, 1979 and 1985, Wheelwright, 1984) proposed that a plant must choose between being a mass producer and being a job shop.

As a mass producer, the company could compete by having low costs, achieved with economies of scale from the high volumes produced. As a job shop, economies of scale were lost, but the company could compete by offering a high customization degree, and the manufacturing system would be extremely flexible.

The increasing of individualized customer requirements (Porter, 1996, Peppers and Rogers, 1997) poses new challenges to the manufacturing system. Customers want customization, speed and low cost simultaneously. In this context, overcoming the trade-off between customization and standardization, delivering timely high volumes of differentiated products at low costs, is a competitive advantage. Companies achieving this competitive advantage early will have superior performance over their competitors.

Mass customization seems to be one possible response to break this trade-off (Feitzinger and Lee, 1997, Tu et al., 2001).
In this text, mass customization will be defined as the capability of offering highly differentiated products, suited individually for each customer’s requirement, with small lead times at low costs (Feitzinger and Lee, 1997).

The Importance of Time to Mass Customization

Early attempts to do any customization at mass producers produced huge increases in complexity and costs. Henry Ford is perhaps the most extreme example: to compete with GM’s variety and switch from the model T, he had to close his River Rouge plant for about one year in 1927 (Sloan, 1965: 192).

Stalk (1988) proposed that the competition had moved from scale to focus and then to variety. This last competition pattern, according to him, requires the competition to be time-based, that is, the faster response player has a competitive weapon on the market battlefield. This responsiveness can be obtained trough flexible Operations, Marketing and R&D.

From the end of 20th Century, though, competition moved again from variety to customization (Lampbell and Mintzberg, 1996). But it has not displaced time-based competition importance. On the contrary, with customization, time increased its strength as a competitive weapon, becoming one of the most important attributes to make customization feasible.

Van Hoek (2001) demonstrated that mass customization and time-based manufacturing practices have positive impacts on the value to the customer, and time-based manufacturing practices have positive impact, themselves, on mass customization.

Competing on Resources

In order to achieve superior performance, a company must develop sustainable competitive advantages (CA) over his competitors (Porter, 1989).

Collis and Montgomery (1995) proposed that, using the Resource-Based View of the firm as a framework, the CA is derived from a unique combination of firm’s resources, subject to the interplay of three external forces: demand, scarcity and appropriability. Resources can be assets, tangible and intangible, and capabilities.

According to Grant (1991), the sustainability of a CA can be assessed by the following attributes: durability, transparency, transferability and replicability. CAs based on assets are less durable then those built on capabilities, because when assets are replaced, changed or lost, firms may have the ability to maintain its capabilities. CAs built over many capabilities and assets coordinated, instead only a few, are harder to be understood by present and potential competitors, and, therefore, harder to be imitated. Grant (1991) called this ability “transparency”. Finally, when a CA is built on capabilities, it requires a very integrated and coordinated set of resources, harder to transfer (buy) or replicate (make) than a single asset.

Customization Strategies

Lampbel and Mintzberg (1996) proposed five customization strategies, according to the Operations stage on which customer can define the product specifications (see Figure 1).

Figure 1: Customization/standardization strategies (Lampbell and Mintzberg, 1996).

Pure standardization means no customization at all: the company pushes products to the market and customers buy them as offered.

Segmented standardization is a higher degree of flexibility, but still the customer does not have direct influence on the final product, he or she has to choose between options the vendor decides to offer. The only influence customer has on segmented standardization is on the distribution process.

Customized Standardization is the first strategy where customer can influence on the final product, by choosing between standardized components. The production remains standardized until the assembly stage, where the customer specifications are met. Lampbel and Mintzberg (1996) stated that, although there will be room in different industries for all five customization strategies, there is a dominant trend towards customized standardization.

Tailored Customization occurs when a customer has the ability to alter the specifications of a product prototype to some extent, during the fabrication stage.

Pure Customization happens when the customer can interfere on the design stage of the product, being able to change every aspect of it (Lampbel and Mintzberg, 1996).

Modularization

Starr (1986:8) defines modularization as “a way to achieve maximum variety through ‘modular’ or ‘combinatory’ productive capacity – that is the capacity to design and fabricate parts that can be combined in numerous ways – and, besides that, the managerial ability to do it”.

Modularization enables mass customization at low costs and must be achieved in three aspects: modular product design, modular process design and a distribution process timely, flexible and not expensive (Feitzinger and Lee, 1997).

Duray et al. (2000), when reviewing the literature about modularity, concluded that mass customization requires modularity to provide unique products in a cost-effective manner by achieving volume-related economies.

Mass Customization Drivers: JIT and Postponement

There are two drivers for mass customization: time-based manufacturing practices (Tu et al., 2001) and postponement (Zinn, 1990; Feitzinger and Lee, 1997; van Hoek, 2001).

Time-based manufacturing practices are derived from Toyota’s JIT system (Ohno, 1997; Shingo, 1996; Monden, 1983), a multidimensional management system supported at manufacturing level by seven elements: shop floor employee involvement in problem solving, reengineering setups, cellular manufacturing, preventive maintenance, quality improvement efforts, dependable suppliers and pull production (Tu et al., 2001; Monden, 1983).

Postponement consists of differentiating a product on the latest possible point at the supply network (Feitzinger and Lee, 1997). This point is called customer order decoupling point (CODP), once it sets the point where the process is decoupled from mass production to customized production (van Hoek, 2001). According to Zinn (1990), there are four types of postponement: labeling, packaging, assembly and fabrication. There is a close relation between these types proposed by Zinn (1990) and the continuum of strategies between standardization and customization proposed by Lampel and Mintzberg (1996). The first two postponement types are related to the segmented standardization strategy, the third is related to customized standardization strategy and the last, to tailored customization strategy.

Case Study

History and intervention methodology:

Vanbro Submersible Pumps Ltd. is as small factory with about 50 employees, located in Southern Brazil. It sells submersible water pumps mainly to the local market and also for other Brazilian states and neighbor countries (Argentina, Uruguay and Paraguay).

The submersible water pump is used to extract water from artesian wells. It consists of two interchangeable modules: an electric engine and a centrifugal multi-stage pump. Four variables are critical when choosing a submersible pump: well diameter (4” or 6”); rate of flow; required pressure and electric power supply available at the field (tension and number of phases).

This leads to the need for a great variety of final products, and there are 582 different variations available in the prices list for December 2001.

The company is privately owned and has two associates, former workers in a local submersible pump company that spun-off their own business, and they are owners-managers. It was founded in 1987 and started as a job shop plant, manufacturing a couple of water pumps per month, made exclusively to order. In 1996, Vanbro launched their first 4” pump model compatible with the NEMA standard. This pump boosted Vanbro’s sales in the following two years to a new level, from a 1996 average of 130 pumps/month to 220 pumps/month in 1998. The average monthly sales in 2001 were 329 water pumps.

In August 2000, Vanbro hired UNISINOS , through EGT (Technology and Management Office), to help them reorganize the plant. The company was growing, and a new building with 300 m2 was being constructed, increasing the 600 m2 area originally available for the plant.

The owners expressed their commitment to their team by means of two requirements: first, no workers would have to be fired due to productivity improvements. The intention was boosting capacity and sales, not reducing costs by job cuts. Second, working at supervisory level, Vanbro had young and motivated undergraduate students. The managers wanted that these employees to dominate the process, so that after the intervention they would be able to walk on their own feet.

Vanbro accepted UNISINOS’s proposal because of its emphasis on capability development, and UNISINOS assigned one professor for the project, that guided his intervention by the action-research methodology. Although the intervention method isn’t he aim of this article, we suggest Thiollent (1997) and Salerno (1999) for a better understanding of it, and Sholtes (1992) for the use of teams and projects.

By that time, the company was not using any just-in-time technique, neither for plant layout, nor for production planning and control (PPC) or continuous improvements in Operations.

First, a working group of 10 people, including the owners, was formed to start studying JIT and layout techniques, to implement them afterwards, in a way suited to their needs. The intervention for the year 2000 was divided in tree parts: in the first two months, there were two weekly meetings with 3 hours each, where basic concepts of Operations Management were presented and discussed, to assess the viability of their implementation in the company. In the next two months, successive versions of the new layout were drawn by tree the members of the group, and there was one weekly meeting with UNISINOS to discuss and improve each layout proposed. The last two months were divided in the implementation supervision and the study of production planning and control techniques, mainly the kanban system for shop floor control.

In 2001 the intervention went on with the first round of improvements in the cells, mainly set-up time reduction, and the study and implementation of kanban for production planning and control of parts and subsets.

Intervention:

Phase One - Diagnosis

The first step started in August 2000, and was a development of internal capabilities, so that Vanbro’s personnel could design all necessary changes themselves. There were classes in Operations Management subjects, readings, discussions and inspections to the shop floor. The main readings were Harmon (1991), Black (1998), Slack (1996), on their chapters about lay out, and Zinn (1990), a paper about postponement. There was also a visit to a large shoe manufacturer using cellular lay out, to show it live at work.

¹UNISINOS is a Brazilian university, located in southern Brazil.

²EGT, an UNISINOS office, maintains university extension activities to the community, such as consulting, training, lectures, etc.

³The assigned professor was Francisco D.C. Ferreira Carmo, one of the authors of this article.

One of the main results of this first phase was a diagnosis that allowed the choice of the operations management techniques better suited for the firm. The most important findings in the diagnosis were:

1) Delivery times were extremely important for customers, mainly in the local market, in two cases. First, when a driller opens an artesian well, it is necessary first to complete the well, because only after that the well is tested the submersible pump can be dimensioned and acquired. In many cases, drilling equipment and crew stands waiting for the pump in the field, so delivery times must be minimal, usually a few hours. Second, when a pump breaks, the owner runs out of water, and will buy the pump that is first delivered, regardless the price. For wholesalers in remote markets, delivery times were still important, but of a few days.

2) The NEMA standards for submersible pumps were extremely convenient, as they forced the design of interchangeable modules. In fact, Vanbro was already using a modular design. According to the owners, this modular design is common in this industry. There were two basic designs for electric engines (4” and 6” diameter); five basic designs for 4” water pumps and four basic designs for 6” water pumps. These eleven frames combined formed the 582 variations quoted above.

3) Vanbro’s purchases were considered small for the buying scales of their suppliers. So, tying stronger partnerships with them was unlikely. Some delivery and quality problems with those vendors didn’t allow JIT delivery of the main materials (cast iron parts, cooper wires and silicon steel), forcing the use of buffer inventories for them. Local wholesalers supplied all other items without major problems. The supply network design, although being a strong element of the JIT system, was considered of low priority by the managers and was not addressed in this intervention.

4) The original layout was completely functional. The main sections were: cutting, drilling, pressing, machining, wiring, assembly and painting. As a consequence, parts had to travel long ways across the factory during fabrication. Nevertheless, an assembly postponement strategy was already being used in an intuitive way, as the assembly section was programmed by customers orders, in opposite to the fabrication sections, programmed by sales forecasts.

5) Production planning and control was extremely informal. In the beginning of the month, a production program was made based on the finished parts inventories and the sales forecast. Fabrication lots were about one month for each part. During the month, as sales demanded, lots were interrupted to fabricate parts needed at the moment. Therefore, there were huge work-in-process inventories all around the shop floor, but only few parts ready to be assembled.

6) No ERP was used to help running the firm, only software for sales, financials, accounting and payroll. The production and purchasing program was made with the help of an electronic spreadsheet with the bill of materials, sales forecasts and inventories. This was feasible due to a very simple product structure and the informality quoted above.

7) As a final result, urgent orders could be shipped within the day, but not without a strike in the shop floor. As production was managed until then, if a different priority scheme was required, all the orders had to be stopped, many setups had to be made and production didn’t flow as expected, with lots of inventories and overtime.

As the capacitating process went on, after many discussions, some points became consensus among the group:

1) A cellular manufacturing lay out could reduce dramatically the traveling of parts around the factory, also making it easier the future implementation of kanban for planning and control.

2) Assembly was the only section programmed in the right way, as it worked “make to order”, postponing the assembly of parts.

3) A limited variety of each part, when assembled, formed a much larger variety of final products. So, manufacturing cells for parts could be programmed and controlled with kanbans.

4) Some acquisitions would be necessary, mainly small auxiliary machines, as painting cabinets, drillers and lathes, to allow the continuous flow of parts in each cell.

The main objective for the project was defined as achieving very short and reliable delivery times without large work-in-process and parts inventories and no finished products inventories.

This would be done in tree stages: First, by reorganizing the lay out, minimizing the traveling of parts around the plant with manufacturing cells. Second, by guaranteeing the availability of parts and subsets for the assembly with the implementation of kanban for the production planning and control, and third by keeping the assembly postponement strategy for the short delivery times.

One major point that made much easier the whole intervention process was the promise from the owners to do not fire anybody after the reorganization. In fact, the number of employees increased from 49 by the end of 1999 to 51 by the end of 2001.

Phase Two – Lay Out Design and Implementation:

The second step started in October 2000, when, in parallel with studies and discussions, tree people from the group started some CAD drawings to show the current lay out and the proposals.

The current situation lay out was analyzed with the help of a simple technique: drawing on the plant the parts flow from raw materials to final products. The result was a landmark in the project, as nobody was really conscious before of how bad the lay out really was, and the urge to improve it was then completely accepted by everybody in the group.

During October and November the team worked first on the definition of the manufacturing cells and after on the lay out design.

The production flow technique (Slack, 1996: 237) for the formation of cells was studied, discussed and applied. One of the main results of using the technique was the learning generated in the group, because during data analysis some auxiliary criteria became necessary.

The criterion for the definition of product families and cells design had to be a mix of factors. Besides process flow, it was considered the raw material type (stainless steel, cast iron and others) and the most adequate lathe for each part, specializing machines that were considered of common use before. This way, product families could be clearly defined.

As a result, the cells and sections for the new lay out were:

1) One cell using the longest lathes for machining stainless steel shafts for the electric engines and the pumps.

2) Two cells using the short lathes for machining cast iron parts. One cell for machining, painting and assembling water pump stages, a high-volume subset, and another cell for machining, drilling painting and assembling all other cast iron parts and subsets, such as bearings, covers, water entrances and water exits.

3) One cell for inserting silicon steel discs to form the magnetic cores for the electric engines cases and shafts, a manual task were piles of steel disks were inserted with the help of a hydraulic press and after cleaned and painted.

4) One cell with mechanical saw and a manual lathe, where steel bars and tubes are cut to form castings and shafts.

The other sections were:

1) High speed press for cutting steel discs. This 80 tons press was one of the main constrains for the new lay out, as it could not be moved without great spending and a long stop in production. So, it remained in the same place. In fact, it was the perfect example of a “monument” as defined by Harmon (1991: 52).

2) Wiring, where cooper wire coils are formed and then inserted in the engine cases. This was a manual job, more similar to a craftwork, and very important for final product quality. Here, engines are differentiated in their electric characteristics (tension and number of phases), in up to six different electric schemes for each case.

3) Final assembly for engines, water pumps and complete sets (engine plus pump). The concept of a final assembly section was fundamental for the assembly postponement strategy. The main improvement in this section was the disposal of shelves for parts and subsets inventories close to the assemblers to ease their work and allow the implementation of kanbans for fabrication. After assembly, engines and sets are tested one by one, and a test record is kept for each product.

4) Final painting, packing and dispatch section.

The three people that made the drawing of the original layout got the task of making the CAD drawings of the new one, and seven successive versions were made and criticized by the group. The method was in fact interactive, respecting some constrains as the press (a “monument”) and the charge and discharge dock for trucks. Cells and sections were placed on the plant and the flow of parts drawn. After, in the weekly meeting the group discussed the result and ways to improve the flow, and a new version drawn for the next meeting.

As a result, the flow of parts on the shop floor was considerably reduced. Besides that, more than 150 m2 of the original building were left available for future expansion and a dock, a dressing room and a ladies room were built. As a matter of fact, the new building wouldn’t be necessary to the capacity expansion planned if the intervention had started before the construction, but it proved helpful because all changes could be done without disrupting operations.

The new lay out was designed in October and November 2000, and during this period a second landmark in the project occurred. While the feasibility of applying cellular manufacturing was still on debate, a pilot cell for machining and assembling water pump stages was made on the shop floor, just moving a few machines in the old building, as the expansion wasn’t ready yet. The success was immediate.

This initiative was completely planned and implemented by the group alone, without the professor’s intervention or previous notice, and it signed that the group was quickly developing the capabilities to design the new system. From this moment on, all the team was convinced of the advantages the new layout would bring, and in two weeks the machines for the shafts machining cell were also moved.

During December 2000 and January 2001 machines were moved to the new building in a sequential way, to avoid disrupting factory operation. After a hold in February, the project continued in March 2001, to make improvements in manufacturing cells and implement kanban for planning and control.

Phase three – Production Planning and Control

In March 2001 the group went back to the textbooks, as more knowledge was necessary to improve the system.

During March and April the group was divided and two subjects were studied in parallel: set-up time reduction and kanban. The weekly meetings were divided, and instead of one long meeting with 3 hours, there were two short ones with one and a half hours.

The references studied for set-up were again Harmon (1991) and Black (1998), on their chapters about set-up, and Shingo (2000), used as a “handbook” of ideas on set-up time reduction.

Set-up time reduction was necessary for the manufacturing cells, and this need would increase as the kanbans were implemented. Although great improvements happened since the beginning, the group stopped meeting after three months, to keep the focus of the project in production planning and control. Besides that, dividing the weekly meetings didn’t prove efficient, as discussions had to stop before reaching main conclusions. Many set-ups have been improved since then with the knowledge and experience acquired, but in an informal way, without an organized team to plan the actions.

The studies for improving the production planning and control and implementing kanban started with Slack (1996), in its chapter on JIT planning and control.

The concept of a mixed system, using MRP for purchases and assembly programming and JIT for internal flow inspired the final design of the system, as seen in Slack (1996: 501). JIT fabrication was implemented with kanban, final assembly remained make-to-order, and purchases remained based on forecasts, exploded though a spreadsheet with the bill of materials and inventories for raw materials.

Other readings on kanban were specific chapters in Black (1998), Tubino (1999) and Schonberger (1984). Again the concept of assembly postponement in Zinn (1990) was discussed and the first trials for implementation started.

Different signs were used as kanbans for different parts: cast iron parts used plastic boxes with fixed cards, placed on shelves that worked as panels. For engine 

shafts, small wood caps that could be fixed on shafts of a panel, and cards for engine cases and engines.

The literature reviewed hasn’t provided clear recommendations on the way to forecast demand to calculate the number of kanbans for each part. The first approach was using average monthly demands divided by four to get weekly demands. But a difficulty emerged during implementation.

The need for extremely fast response at the final assembly showed that, as peak demands in one day could use the average weekly demand for a part, parts shortages could happen. System was then recalculated using the peak weekly demands in the last year for each part. It implied in fact an increase in parts inventories, but the preference was to assure the availability of parts for final assembly, and work-in-process had already been considerably reduced.

The kanban system developed addressed the problem of lack of a MRP II software by eliminating the need of MRP II. Kanbans themselves allowed decentralized production programming by the workers at the shop floor, eliminating the need of a centralized planning for parts fabrication, just for assembly, as will be shown later on this text.

The production planning and control system designed for the firm is shown on Figure 2 below.

Figure 2: Vanbro Production Planning and Control Flowchart. Source: the authors.

The choice of the customer order decoupling point (CODP) would force the planning to switch from kanban to customer orders. The system was, from the early stages of the intervention, conceived as an assembly postponement strategy. However, process peculiarities of water pumps production have indicated the existence of more than one CODP, problem not addressed by the literature reviewed.

Water pumps are assembled exclusively make-to-order, and CODP is in assembly section. There are five basic kinds of 4” pumps, according to flow rate. Pressure is increased just adding more stages, in a total of 39 options. For 6” pumps, there are four basic kinds, and 69 options. The variability increase occurs only at water pump assembly.

For engines the situation is much more complex. There are only two basic frames for engines, 4” and 6”. There are 10 power options for 4” engines (0,5 HP up to 6 HP), and 20 power options for 6” engines (1 HP up to 18 HP), and almost all components are common in each frame. The only change is on engine’s length, as power is increased extending the magnetic core.

Although frames are simple, each core can have up to six electric assemblies, executed in the electric wiring section, where the type of wire, coil’s size and linkage can change, generating 105 options for 6” engines and 57 options for 4”, a total of 162 types. After the wires are placed, the engine cannot be changed, so CODP for engines could not be in final assembly, but in the case wiring section, at the beginning of the process.

A sales ABC curve analysis (Schroeder, 1985: 411) of these 162 options was used to determine the planning strategy for each model. “A” and “B” engines would be programmed by kanbans and “C” engines directly by customer orders, similar to Slack’s (1996: 501) proposal of JIT for common items and MRP for special ones. Having the “A” and “B” engines ready to pump set assembly allows an urgent customer order to be shipped within 2 to 3 hours, without flow disruption at the shop floor. The engine was then turned into a subset for the pump production.

The new procedure designed for programming the assembly and wiring sections was: first, the programmer would check the availability of assembled engines for the order. If there are engines ready, he would reserve these engines, removing their kanban cards and sending them to the case wiring section for replacement. If the engines for the order were not available, or they were “C” engines, fabrication orders for them would be placed at case wiring section.

In the case wiring section, coils are made and inserted in engine cores in the same sequence they are going to be used at the assembly section, assuring the smallest lead times possible.

After implementing the new system, 80% of orders programmed for the day were ready for dispatch before lunch. During the afternoon, the assembly worked on the remaining orders and assembling engines to refill the kanban. The dependable 2 to 3 hours lead times for urgent orders is far better than the industry practices, while bigger companies deliver those urgent orders within a couple of days.

Conclusion

Examples of mass customization in the literature are normally on companies with high technology end-user products, like HP and Nike, and deploy a large amount of Information Technologies (IT) resources to make the process viable. Our study showed a case on a very small company, with a product often “invisible” to their end-users, except when broken. This company was not using high technology neither for product, process, supply network or IT. These technologies were not necessary to the mass customization deployment.

One insight was on the effect of time-based competition: the effort to compress delivery times to few hours turned the manufacturing very similar to a service operation, like a fast-food, where demand peeks are measured in hours or even minutes, instead of seasons, months or days, formerly usual in manufacturing operations. This led to a change in the demands forecast used to dimension the kanbans from average to peeks.

The main improvement on the organization studied was on capabilities, mainly knowledge, not on assets. Mass customization was not enabled by faster parts machining or drilling, but by better resource usage. Few investments were necessary, at points where the original assets could not solve the problems, confirming the statement of Shingo (1996: 171) on “catalog engineers”. The machines purchased were merely small painting cabinets, small drillers, standard plastic boxes and shelves to store parts in the kanbans at the cells.

Like Toyota’s manufacturing system (Ohno, 1997), Vanbro’s system was constructed internally based on its capabilities and customers’ demands, not directly transferred from another company where it had succeeded before.

This case also showed the importance of employee involvement on the design of the production system. Early involvement had the following advantages: the system design capitalized on the large operational and market experience from both the owners and workers; the risk of “sabotage” was very low, because, on employees’ words, it was “our system”.

Questions for Future Research

This case study showed evidences that postponement and JIT seems to contribute to mass customization. Due to the generalization problems intrinsic to this methodology, and the absence of empirical tests on the literature reviewed, further research on this issue could explain the relationship and the combined effects of postponement and JIT (or time-based manufacturing practices) on mass customization.

The relationship between IT and mass customization was not addressed empirically in the literature reviewed. As our study showed, mass customization was possible with no investment on IT. Further research would show criteria for deploying and investing on IT to enable mass customization.

The existence of more than one CODP in a single production process was not addressed on the literature reviewed. The absence of a theoretical framework to guide the decision on the placement of the many CODP found in the water pump factory was guided by intuition. CODPs were placed in the point from where the variety would grow exponentially. Further research could address decision criteria for multiple CODP placements inside the process.

References

BLACK, J. T. O projeto da fábrica com futuro. Porto Alegre: Bookman, 1998.

COLLIS, David J. and Montgomery, Cynthia A. Competing on Resources: Strategy in the 1990s. Harvard Business Review. Boston. July-August 1995. pp 118-128.

DURAY, Rebecca; WARD, Peter T.; MILLIGAN, Glenn W. and BERRY, William L. Approaches to mass customization: configurations and empirical validation. Journal of Operations Management. Volume 18, Issue 6 November 2000 Pages 605-625.

FEITZINGER, Edward; LEE, H. Hau. Mass Customization at Hewlett-Packard: The Power of Postponement. Harvard Business Review, pp. 116–121, January-February, 1997.

GRANT, Robert M. The Resource-Based Theory of Competitive Advantage: Implications for Strategy Formulation.California Management Review. Spring 1991, pp. 114-135.

HARMON, Roy ; PETERSON, Leroy D. Reinventando a fábrica. Rio de Janeiro: Campus; 1991.

HAYES, Robert H; WHEELWRIGHT, Steven C. Linking manufacturing process and product life cycles. Harvard Business Review, pp. 133–140, January-February, 1979.

WHEELWRIGHT, R., HAYES, R. H. Competing through manufacturing. Harvard Business Review, [s. l.], v. 63, n. 1, p. 99 - 109, January-February 1985.

LAMPBEL, Joseph; MINTZBERG, Henry. Customizing Customization. Sloan Management Review, pp. 21–30. Fall, 1996.

MONDEN, Hiasuhiro. Toyota Management System – Linking the Seven Key Functional Areas. Portland: Productivity Press, 1993.

OHNO, Taiichi. O Sistema Toyota de Produção: Além da Produção em Larga Escala. Porto Alegre: Bookman, 1997.

PEPPERS, Don; ROGERS, Martha.E. The One to One Future: Building Relationships One Customer at a Time. New York: Currency Doubleday, 1997.

PORTER, Michael E. A Vantagem Competitiva. Rio de Janeiro: Campus, 1989.

PORTER, Michael E. What is strategy? Harvard Business Review. 74 6, pp. 61-78, 1996.

SALERNO, Mario S. Projeto de organizações integradas e flexíveis. São Paulo: Atlas, 1999.

SCHOLTES, Peter R. Times da qualidade. Rio de Janeiro, Qualitymark, 1992.

SCHONBERGER, Richard J. Técnicas industriais japonesas. 3a ed. São Paulo: Pioneira, 1984.

SCHROEDER, Roger G. Operations Management. 2nd ed. New York: McGraw Hill, 1985.

SHINGO, Shigeo. Sistemas de Produção com Estoque Zero. Porto Alegre: Bookman, 1996.

SHINGO, Shigeo. Sistema de troca rápida de ferramentas. Porto Alegre: Bookman, 2000.

SKINNER, W. Manufacturing – Missing link in corporate strategy. Harvard Business Review, Boston, Estados Unidos, v. 47, n. 3, p. 5 - 14, May-June 1969.

SLACK, Nigel et alli. Administração da produção. São Paulo: Atlas, 1996.

SLOAN, Alfred P. Minha vida na General Motors. Rio de Janeiro: Record, 1965.

STALK, G.JR.. Time – The Next Source of Competitive Advantage. Harvard Business Review, Boston, p. 41 - 51, July-August 1988.

STARR, Martin K. Produção modular – Um novo conceito. In Coleção Harvard de Administração. Vol. 11. São Paulo: Nova Cultural; 1986; pp 7 - 30.

THIOLLENT, Michel. Pesquisa-ação nas organizações. São Paulo: Atlas, 1997.

TU; Qiang; VONDEREMBSE, Mark A. and RAGU-NATHAN, T. S., The impact of time-based manufacturing practices on mass customization and value to customer, in Journal of Operations Management, Volume 19, Issue 2, February 2001, Pages 201-217.

TUBINO, Dalvio Ferrari. Sistemas de produção: a produtividade no chão da fábrica. Porto Alegre: Bookman, 1999.

van HOEK, R. I. The rediscovery of postponement a literature review and directions for research. Journal of Operations Management, Volume 19, Issue 2, February 2001, Pages 161-184.

WHEELWRIGHT, S.C. Manufacturing strategy: defining the missing link. Strategic Management Journal, Chichester, v. 5, n. 1, p. 77 - 91, January-March 1984.

ZINN, Walter. O Retardamento da montagem final de produtos como estratégia de marketing e distribuição. In: RAE - Revista de Administração de Empresas. São Paulo, vol. 30, no 4; pp 53-59, out-dez 1990.