Disruptive Innovation: A Case of Full Mold Casting

1. IntroductionThere exist sustaining innovations that improve product performance according to existing criteria. In addition, there exist disruptive innovations that can replace existing products, first improving product performance in a different set of criteria despite lower performance in the original criteria, and subsequently causing improvements according to existing product performance criteria (Christensen, 1997).Christensen and Raynor (2003) assumed two types of customers with regard to disruptive innovation: entirely new nonconsumers in markets outside of existing markets, and overshot customers in existing markets.1 New-market disruptive innovations provide new value to nonconsumers through innovations for new markets that subsequently cause disruptive innovations in existing markets. Low-end disruptive innovations provide products and services that grant value in line with low-end overshot customers, but subsequently target customers from the high end (Christensen & Raynor, 2003). Keywords for disruptive innovation are cost, traditional performance, and ancillary performance change (Utterback & Acee, 2005).2 While new-market disruptive innovations may lag in traditional performance, they acquire new markets with innovations in superior ancillary performance and permeate from new markets to existing markets. On the other hand, low-end disruptive innovations exhibit poor traditional performance; nevertheless, innovations with lower costs enable them to withdraw customers from existing markets. 3 New-market disruptive innovations can also be used to replace products on the low end of existing markets; thus, conceptually, there is continuity between new-market and low-end disruptive innovations (Christensen & Raynor, 2003).Existing companies can respond to a sustaining innovation even when the innovation is radical; however, with disruptive innovation, existing companies often fail to form a response. This is because existing companies are bound by their existing value network in which they have developed a customer base and business relationships (Christensen & Rosenbloom, 1995). As understood through the value network, when a value network cannot respond to nonconsumers that demand new value, new-market disruptive innovations occur; when low-end customers are taken from a primary value network with existing products and services, the result is low-end disruptive innovation (Christensen & Raynor, 2003).Can we then assume that there are only two customer types, nonconsumers and overshot customers, with regard to customers who influence the value of a disruptive innovation?4 In Japan's large-scale casting industry, the full mold casting (FMC) method used by Kimura Chuzosho Co., Ltd. became a disruptive innovation that acquired customers with specific requirements unmet by existing wood pattern-based sand mold casting. These customers were neither nonconsumers nor overshot customers.2. Process Innovation of FMCAs the name implies, a wood pattern is a pattern made from wood. Sand mold casting in the casting industry has extensively used these wood patterns. The wood patterns are filled with molding sand, and a mold is formed by solidifying the molding sand. The space created by the removal of the wood pattern is filled with melted metal, and a casting is formed. At present, medium- and small-scale castings are often made by using patterns of metal or plastic, while molds are often formed with metal. However, even today, large-scale castings are made by using sand mold casting with wood patterns.In contrast, FMC also uses molding sand, though it uses patterns made from foam polystyrene instead of wood patterns. In FMC, the pattern fills the interior of the mold, and subsequently melted metal replaces the pattern inside the mold, thereby forming a casting.The process of wood pattern-based sand mold casting is as follows. 1) A wood pattern is formed (this pattern is a model of the finished product). …


Introduction
There exist sustaining innovations that improve product performance according to existing criteria. In addition, there exist disruptive innovations that can replace existing products, first improving product performance in a different set of criteria despite lower performance in the original criteria, and subsequently causing improvements according to existing product performance criteria (Christensen, 1997). Christensen and Raynor (2003) assumed two types of customers with regard to disruptive innovation: entirely new nonconsumers in markets outside of existing markets, and overshot customers in existing markets. 1 New-market disruptive innovations provide new value to nonconsumers through innovations for new markets that subsequently cause disruptive innovations in existing markets.
Low-end disruptive innovations provide products and services that grant value in line with low-end overshot customers, but subsequently target customers from the high end (Christensen & Raynor, 2003). Keywords for disruptive innovation are cost, traditional performance, and ancillary performance change (Utterback & Acee, 2005). 2 While new-market disruptive innovations may lag in traditional performance, they acquire new markets with innovations in superior ancillary performance and permeate from new markets to existing markets. On the other hand, low-end disruptive innovations exhibit poor traditional performance; nevertheless, innovations with lower costs enable them to withdraw customers from existing markets. 3 New-market disruptive innovations can also be used to replace products on the low end of existing markets; thus, conceptually, there is continuity between new-market and low-end disruptive innovations (Christensen & Raynor, 2003).
Existing companies can respond to a sustaining innovation even when the innovation is radical; however, with disruptive innovation, existing companies often fail to form a response. This is because existing companies are bound by their existing value network in which they have developed a customer base and business 1 In addition to nonconsumers and overshot customers, Christensen, Anthony, and Roth (2004) also assume the existence of undershot customers. These are customers among an existing customer base that are unsatisfied by current products. They make it easier to generate sustaining innovations. 2 Utterback and Acee (2005) noted the necessity of seeing eight patterns of change by various combinations of increases or decreases in cost, traditional performance, and ancillary performance. 3 Customers value them as having not only lower cost but also being simpler. relationships (Christensen & Rosenbloom, 1995). As understood through the value network, when a value network cannot respond to nonconsumers that demand new value, new-market disruptive innovations occur; when low-end customers are taken from a primary value network with existing products and services, the result is low-end disruptive innovation (Christensen & Raynor, 2003).
Can we then assume that there are only two customer types, nonconsumers and overshot customers, with regard to customers who influence the value of a disruptive innovation? 4 In Japan's large-scale casting industry, the full mold casting (FMC) method used by Kimura Chuzosho Co., Ltd. became a disruptive innovation that acquired customers with specific requirements unmet by existing wood pattern-based sand mold casting. These customers were neither nonconsumers nor overshot customers.

Process Innovation of FMC
As the name implies, a wood pattern is a pattern made from wood.
Sand mold casting in the casting industry has extensively used these wood patterns. The wood patterns are filled with molding sand, and a mold is formed by solidifying the molding sand. The space created by the removal of the wood pattern is filled with melted metal, and a casting is formed. At present, medium-and small-scale castings are often made by using patterns of metal or plastic, while molds are 4 Some critics state that, from a technology perspective, it is unclear what is disruptive innovation and at the point at which an innovation becomes disruptive, making it possible to declare something disruptive after the fact (Danneels, 2004). The S-shaped technology curve has strong randomness, and existing technology cannot simply be substituted for disruptive technology. Thus, the view of existing vs. disruptive technologies is an oversimplification (Sood & Tellis, 2005;Tellis, 2006). When we swap one metric for another, in some cases we can see that there are no differences between existing and disruptive technologies (Takahashi, Shintaku, & Ohkawa, 2013).
often formed with metal. However, even today, large-scale castings are made by using sand mold casting with wood patterns.
In contrast, FMC also uses molding sand, though it uses patterns made from foam polystyrene instead of wood patterns. In FMC, the pattern fills the interior of the mold, and subsequently melted metal replaces the pattern inside the mold, thereby forming a casting.
The process of wood pattern-based sand mold casting is as follows.
1) A wood pattern is formed (this pattern is a model of the finished product).
2) The top and bottom of a mold are prepared.
3) The wood pattern is placed in the mold and filled with sand. 4) As the sand solidifies, the wood pattern is removed. 5) The surface of the mold is coated with a slury of carbon, allowed to dry, and thereafter top and bottom portions are prepared. 6) The core of the mold, equivalent to the hollow portion of the casting, is then prepared. 7) The core is placed into the bottom part, and the top part is then combined. 8) The completed mold is filled with melted metal. 9) The casting is allowed to cool and solidify. 10) The casting is removed, processed, and polished.
In contrast, the process flow used by FMC is as follows: 1) A pattern is made from foam polystyrene.
2) The surface of the pattern is coated with a slury of carbon and allowed to dry.
3) The pattern is placed inside the mold and filled with sand. 4) When the sand solidifies, the mold is filled with melted metal. 5) The casting is allowed to cool and solidify. 6) The casting is removed, processed, and polished.
Sand mold casting using wood patterns has been in existence for a long time, while the basic patents for FMC were granted in 1958 in the US, making it a new method of casting. In comparison with wood pattern-based sand mold casting, FMC is a process innovation that obviates the need to separate top and bottom, the removal of patterns, and cores (Kanno, 2004). Because FMC does not separate top and bottom, it does not have a burr at the joint in the final product.
Because FMC does not require the removal of the pattern, there is no need for designs to consider the manner in which patterns are to be removed. Because there is no core, designing complex shapes is greatly simplified. Creating such complex shapes using wood pattern-based sand mold casting required highly developed skills; however, these are not necessary with FMC.
According to Ozerov, Shuliak, and Plotnikov (1971), the per capita productivity for simple casting by FMC improved approximately three times that for wood-pattern based sand mold casting. The cost to produce patterns equivalent to 22-ton castings was approximately $17,000 using wood pattern, while only $1,700 using foam polystyrene patterns. In one casting shop for an automobile factory, the implementation of FMC resulted in a 30% reduction in pattern-related labor and a 25% reduction in model-related labor.
On the other hand, FMC does have its problems. Wood patterns in sand mold casting can be used almost permanently once created, and this is important for mass production. In FMC, however, melted metal is poured directly onto the pattern, making the pattern usable only once. This is not a problem for prototypes and single casting, but leads to issues with uniformity and cost in mass production. In addition, the surface of the mold is adversely affected by gasses released when the foam polystyrene is utilized by the melted metal.
This is a major problem with melted metal permeating into the mold in different ways compared with traditional methods, and causing more scrap. Although FMC is in principle superior, scrap leads to excessive costs.
As a result, many companies showed an interest in FMC, but did not find it marketable, and thus their interest waned. However, Kimura Chuzosho Co., Ltd is one company that has used FMC and been successful in replacing wood pattern-based sand mold casting in the marketplace.
The casting of machine tools often involves mass production, and the production volume share of Kimura Chuzosho in that market was 10% in 2000, exceeding 20% in 2005. 5 Despite conventional wisdom proclaiming that FMC is difficult to use, excluding large-scale casting, Kimura has expanded their market share in that area. Their success can be attributed to two points: their accumulation of innovations and the use of IT. We describe both of these below.

Achieving Mass Production
Computer aided design (CAD)/computer aided manufacturing (CAM) systems support design and manufacturing, and became popular in the manufacturing industry in the 1980s. In 1987, Kimura Chuzosho developed a CAD/CAM system that enabled the company to create CAM data from user-generated CAD data, and subsequently create patterns using NC machines and blocks of foam polystyrene. This enabled the company to begin mass production of patterns and to develop mass production for machine tool casting.
Nevertheless, the yield at the time was only approximately 20%. First, only approximately 30% of patterns were produced internally, and limits on computer processing power meant a great deal of manual work. This manual work required ten years of experience, and there was often a great disparity in the competence of workers. 9 Kimura equipped itself with engineering workstations in 1995, and purchased the CAD/CAM package "CADCEUS" in 1996. However, calculations that today only take one minute at the time took two hours. Even so, the company invested excess capacity into the mass production of machine tools in an effort to expand production in 1995. Simultaneously, it established a new factory in Gunma to expand its pattern production capabilities.
Kimura's production volume share of the metal processing machinery casting market rose from 0.3% in 1987 to 4.0% in 1996.
Its share of the press mold casting market rose to 54% by 1996.
Soon, CAD began to be widely used among Kimura's customers, and the company began accepting CAD data from them. 10 In doing so, Kimura headed even further in the direction of digital production of patterns. The company aimed for the full mechanization of pattern production in the Omaezaki Number 2 Factory in 1997, and at the time of plant startup it purchased four NC machines (bringing to a total of 19 in use through the Kimura Group), and also developed the "Pro/E" CAD/CAM software package.
In 1998, Kimura jointly developed automated CAM software with a third party software manufacturer, fully automating a process that until that time comprised manual design, calculation, and decision making by humans. In 1999, PC-based CAD/CAM software appeared in the market promising both high performance and low cost, and 9 IT is implemented because it was clear that it reduced cost and improved quality (Ichikohji, 2013). In the case of FMC, these benefits were unclear in the beginning. 10 The sharing of CAD data between corporations reduces the time it takes to solve problem, while also having the negative effect of increasing coordination between the same companies. It does not always result in a net benefit (Ku, 2004 This series of efforts improved accuracy from 5 mm to 0-1.5 mm, and CAD/CAM usage increased from 30% in 1999 to 100% in 2002. The company was successful in fully mechanizing the production of patterns via NC machining using three-dimensional (3D) solid CAD data. This mechanization enabled the company to withdraw from pattern creation that relies on individual skill. Furthermore, work that required 10 years of experience could now be performed by CAD operators with only one year of experience.
The dramatic rise in pattern-related productivity meant that the number of days to create patterns for press mold side panels was reduced from 20 days in 1996 to only seven days in 2012.
Mechanizing pattern production allowed part-timers rather than skilled professionals to be responsible for the work, which helped reduce costs. Manufacturing lead times using sand mold casting with wood patterns required three months from design to actual casting, but Kimura was in the end able to reduce this lead time to approximately three weeks.
Even sand mold casting with wood patterns is now performed by using 3D CAD to design wood patterns and CAM software to produce them. However, 3D CAD/CAM has the benefit of enabling mass production with high pattern accuracy, and as the technology has also enabled mass production via FMC. The use of 3D CAD/CAM in the pattern process, as well as casting innovations, has improved quality and facilitated mass production. Kimura's production volume in the machine tool casting market rose to 23% in 2010.

Disruptive Innovation: A Case of Full Mold Casting
Disruptive innovations may occur due to reduced product performance according to existing criteria, while they improve performance in other criteria, and end up as alternatives to existing products by improving performance according to existing criteria as well (Christensen, 1997).
In the large-scale casting market, the basic criterion of customers is mold surface quality. Without a certain level of quality, the casting cannot serve its purpose; however, achieving the requisite level of quality can cause cost issues. As opposed to wood pattern-based sand mold casting that achieve the sufficient mold surface quality, FMC suffered from poor quality. However, some customers welcomed shorter lead times despite lower quality because automobile mold companies could adjust surface quality of castings. FMC acquired customers who prioritized shorter lead times as their primary criteria, and thereafter FMC improved casting quality and productivity. This led to FMC's success in the automobile mold casting market.
Furthermore, after success in the single machine tool casting market, which required higher mid-range casting surface quality, Kimura Chuzosho successfully began mass producing with shortened lead times and improved quality and accuracy for high-end, mass produced machine tool castings and the company won a higher market share. Thus, FMC is a disruptive innovation against wood pattern-based sand mold casting.
Did FMC allow for the acquisition of new, nonconsumer customers? Production for automobile mold casting and machine tool casting already occurred using wood-pattern based sand mold casting, and this production method had existing customers. Thus, FMC did not acquire nonconsumers. Were customers of automobile mold casting and machine tool casting, or customers for mass-produced castings for machine tools considered overshot customers? Developing low-end disruptive innovations is a low-cost business model (Christensen & Raynor, 2003). As to automobile mold casting, automakers were fine with lower mold surface quality in return for shortened lead times, but customers did not approve lower quality in exchange for reduced cost. For machine tool castings, machine tool manufacturers do not adjust mold surface quality, and they demand high quality rather than shorter lead times. Moreover, mass production requires higher mold surface quality as well as replicable accuracy and production lead times, which are short enough to support mass production. In Japan's large-scale casting industry, FMC became a disruptive innovation that withdrew customers from wood-pattern based sand mold casting, and these customers had specific demands but were neither nonconsumers nor overshot customers.
The logic behind disruptive innovation is far simpler than the customer analyses provided by Christensen et al. (2004). Even if a company can acquire only a small portion of customers with particular, specific requirements using quality, cost, delivery (QCD)-related technologies, which are perhaps even toy-like in comparison with existing technologies, such company will have opportunities to improve overall QCD as its business continues.