STRATEGY AN ANALYSIS OF SUB-CONTRACTING RELATIONSHIPS BASED ON THE SUB-CONTRACTOR/CUSTOMER TECHNOLOGY EXCHANGE PORTFOLIO: SOME EMPIRICAL FINDINGS* Corrado lo Storto ODDISSEO-University of Naples and DIMEG-University of Padua ABSTRACT In the last decade, subcontracn'ng relationships have assumed a particular weight in the strategy ofmanagers searching for ~iciency flexi bi li ty, and quali ty. Subcon traciing is not a new industrial practice. Reasons and waysfirms'ecourse toit have changed. Exchange relation- ships have become highly complex in their nature. In order to manage complexity, multiple channels for the coordinanon of the acnvities are developed, and an intensive exchange of technology is established. Tlus paper reports some findi ngs of an empirical exploratory study aimed at investigating the nature of technology flow in the customer firmlsubcoturactor interaction. A sample of Italian small and medium subcontractors was considered. INTRODUCTION In the last few years, collaborative agreements between entrepreneurial firms and large companies have become increasingly common, Subcontracting is a frequent form of cooperation in which a small- or medium-sized firm produces parts for a large final manufacturer or performs operations for that producer as a wage work (Lazerson, 1990). These relationships exist in almost every industrial sector. They are common in industries characterized by complex products, some of which involve thousands of components (i.e., automobile, aircraft and tool machine manufac- turing). By using a pool of external suppliers a greater flexibility can be combined with a constant high productivity. Sharing production among firms determines a division of technological areas of specialization, and the subcontracting system becomes a distribution channel for technical knowledge, routines and procedures transfer, and organizational learning for small firms. v This paper received ihc Distinguished paper Award ca the I994 SBIDA National Confcvcncein San Antonio. It nws noi vc va wed by thc ISBS Editorial Advisory Beard. Tlu author wishes io achnowlcdgc ihc financial suppon gi van by the CNR funds of Psog coo Strategic o "Tvosfcrimcnlo dcllc iccnologw dci psogeai finalesati", research unit of Naples, that made u possible io collect the data. Ftothcrly, hc is indebted io Prof Maria Raga and Pr% Giuseppe Zollo for their helpful suggestions during ihe preparation of this paper. Nccdlcss to say, he u very grateful for thc precious commems and suggestions made by an anonymous reviewer. 53 Considering the foregoing discussion, an interesting research question arises. That is, "What kind of relationship exists between the nature of thc technological flow of the customer firm or subcontractor business exchange, and product and subcontractor's characteristics?" AN OPERATIONAL CONCEPT OF TECHNOLOGY A flirm's technology is the set of technologies that it utilizes to carry on its activities. This set of technological competencies and capabilities continuously modifies as the result of efforts of internal or external development. Technology can be embedded in the individuals, in the structure and culture of organizations, in its records and rouunes, or in its equipment (Hayes & Wheelwright, 1984). Diverse modes of organizational learning convert individual expertise and knowledge into organizauonal technologies. According to Nelson and Winter (1982), firms develop routines that reflect the organization's ability to adapt to changing circumstances and establish distinctive capabilities and skills. Unfortunately, technology sometimes remains an ambiguous concept, difficult to opcrationalize. A way to overcome this difficulty is to break it down to its basic components. In this paper, a definiuon of technology adapted from that proposed by the Technology Atlas Team (1987) is used. Four basic components of technology are identified: hardware (HW), object- embodied technology; human skill (HS), people-embodied technology, which is the know how of carrying out tasks; information (IN), document-embodied technology; rules and procedures (RP), institution embodied technology. This latter establishes the nature and sequence of the information flows, machines to be used, and skills required. Table I reports the classification used. This taxonomy allows the analysis of how technology flows among firms and the ways a subcontractor can internalize new technology. Table I The Four Components of Technology Technology hardware-embodied (HW) It is in form of equipment, machines, tools. Technology document-embodied (IN) It is in form of manuals, designs, schemes. It describes how to use a single object. It is always codified. Technology people-embodied (HS) lt is in foun of judgements, opinions, sugges- tions, impressions, .;kills. It is not codifiable and transferable only through people. Technology rule-procedure embodied (RP) It is in form of scheduling plans, procedures, organizational rule«. It establishes what kind of relations exist among HW, IN, HS for task ex- ecution. It can be formalized and codifiable with a different degree. 54 TECHNOLOGY TRANSFER& ADAPTATION& AND COORDINATION MECHANISMS Subcontractors and customers often establish lasting collaborative relationships, and when the business exchange accounts for a relevant share of subcontractor sales or customer's needs, a significant adaptation process occurs (Hallen, Johanson, /k Seyed-Mohamed, 1991). This adaptation develops both as a process of organizational fit and as a process of day-by-day coordination. The process of fit between customer and subcontractor is necessary to establish a common language to make effective communication between the two parties. It allows an increase in the technological capability of one of the parties involved, and to make it homogeneous and compatible with that of the other. This adaptation usually requires a substantial modification for the organizations involved, and considerable invesunents, so to increase dependence between the parties. The subcontractor can adapt its manufacturing process to that of its customer by purchasing equipment which can use part programs for Numerically Controlled (NC)'achines created by the customer, by introducing logistic systems, common scheduling and planning procedures, and routines for quality control. TI&e other form of adaptation almost occurs on a day- by-day base, with the coordination of the subcontractor's operations with those of its customer. Coordination is necessary to cope with the uncertainty, ambiguity, and inter-dependence of the system, arising from the absence of information or the presence of conllicting interpretations of a situation, and the organization of producuve system (Galbraith, 1973; Thompson, 1967, /k Weick, 1979). In order to reduce uncertainty and ambiguity and cope with the system inter- dependence, organizations have to transfer and process information (Daft k. Lengel, 1986). The adaptauon and coordination mechanisms represent effective means for the transfer of technology across fums. Empirical research aimed at studying the interaction between small and large firms identified 11 channels frequently used by firms to transfer technology (Esposito rk Raffa, 1991): raw materials, sub-components and pre-machined parts, machinery and equipment, training, advice on specific issues, in-progress check, advice for the adoption of quality rouunes, quality management procedures, meetings, documents, and different forms of collaboration at the start of the transaction. These channels transfer technology under the four forms described. METHODOLOGY Data were collected through face-to-face interviews with entrepreneurs, technicians, and plant managers I'rom the end of 1990to 1991. According to the author's experience, a structured face-to-face interview approach can yield more reliable data. It allows the interviewer to elaborate on those questions that might cause difficulty, and to gain a visual impression of the object of the survey. Each respondent was interviewed by means of a questionnaire seeking data on economic and other firm characteristics, such as location, sales, employment, technical/manufacturing capability, and interaction modes with the main customers. A sample of 59 subcontracting firms was studied in depth. All firms investigated are located in Italy. They subcontract to customers selling in the aerospace/electronics, automotive, and medium-low technology industries (electrotechnology, tool machine building, texule-machine). NC is a sysrem fa'iu.'uromarion of sorio hie «ork phases on mar or working maehi ru'roots. 55 The average size of a sample's firm is 82.1 employees (sd = 112.7),sales per year 13.234billions of Lira (sd = 29.37). Every subconuacting uansaction is characterized by a certain combination of the four technology components. This combination determines the technological portfolio of the customer/subcontractor exchange. The model of the technology transfer channels proposed by Esposito and Raffa (1991)is used here to study the nature of technology transfer from a customer to its subcontractor, and identify the technological portfolio. Each respondent was asked to judge the importance and utility of each channel to uansfer technology into his factory on a value scale from I to 9. According to Saaty (1980), this scale captures fairly well the preferences of an individual. Every channel vehicles technology in the form of hardware, skills, information, and organizational rules and procedures at a different rate. In order to measure how each channel affects every component, a major assumption was introduced —that the influence of thc flow of technology along the channel on each technology component was given by the following relation: TCi = IS (wlj Cj + wllj Cj+..I/Vmaxi j= I,...,ll where TCi is the generic technology component Cj is the utility-value given to channel j by the respondent Vmaxi is the maximum possible value for the channels influencing that component wj is a weight measuring the magnitude of the effect of that channel on the component i wlj measures 1st order effects, wllj measures 2nd order e(fccts, and so on. Some discussion with business consultants, entrepreneurs, and firms managers made it possible to build a matrix of weights accounting for the contribution of the channels of technology transfer to each component of technology. A "pairwise comparisons" approach common in the Analytic Hierarchy Process (AHP) technique was used to determine weights.' second assumption was done, that all the effects were of the first order. Consequently, only wlj values were considered. The mapping of the relational system for the evaluation of the technological portfolio is reported in the Appendix. Cluster analysis was used to group firms, using as the clustering variables the four components of technology. The K-Means algorithm proposed 1&y MacQueen was used as clustering technique. Because the algorithm required to specify in advance the number of clusters (K), the software program was made rerun for various choices of K in order to obtain K not repetitive combinations of technology component measuremcnts having the highest F-values. For every subcontractor, its first and second customers were considered. On average the sales for the first two customers account for more than 65 percent of total sales. In order to take into account the volume of the business exchange, the flow of technology from the customers was weighed with the percentage of sales. Analysis of Variance (ANOVA) was used to test the difference between the four clusters. ltHP is o tcchniquc devidoped to find priori ties of dilfer cut alter noti vas r chai ve to an oh/ cct reducing a scq nenes of prioruy prohu mein to a scquc net ofpairwisc comparisons according to certain cri teria (l062 ln this case tlw criteria werc thc foiu technology cornponcnis, thc el ternati ves the I I channels of techno!ogy transfer. 56 RESEARCH FINDINGS Table 2 shows the correlation matrix among the values of technology components for the main two customers. A compensation effect between some components of technology emerges. Particularly, an increase of HW determines a decrease of IN and, with a lower intensity, a decrease of HS. Generally, the more technology is transferred lending specific equipment and specialized machinery to the subcontractor, or providing it with the raw materials, the less it is necessary to transfer a rich written technical documentation. As an illustration, let us think of a firm that receives a die and the raw material to make plastic moldings flom the customer. This firm will need only some indications about how to blend the plastic powders, the value of temperature, and thenumberofpiecestoproduce. Thesameisfor people-embodiedtechnology. Thestrongeris the weight of the physical component of technology, the less relevant is the weight of people. In the same way, an increase in the value of HS determines a decrease in RP (and vice versa). This is particularly true when organizational rules and procedures are easy to be codified, and after being transferred into a firm, they become routines. That is the case of the procedures a fum must implement to manage quality control. This compensation effect among the components of technology remains considering the relationship with each customer separately. The outcomes from cluster analysis confirm this compensauon effect. Table 2 Correlation Matrix of the Technology Components —First and Second Customer HW IN HS RP HW 1.000 (0.000) IN -0.734 1.000 (0.000) (0.000) HS - 0.542 0.148 1.000 (0.000) (0.263) (0.000) RP -0.313 0.1&4 -0.522 1.000 (0.016) (0.162) (0.000) (0.000) Source: database ODISSEO Note: probabilities are in brackets Banlett-Chi Square Statistic: 621.018 Prob= 0.000 Table 3 illustrates what happens to the four technology components passing from the first to the second customer. The variables considered in the correlation were obtained as the difference between the values of each technology component for the first two customers, then normalized by dividing it for the maximum value of the technology component. Technology components tend to vary in the same direction. An increase in the flow of technology in one of the four forms determines an increase in the flow in the other three forms. 57 Table 3 Correlation Matrix ofthe Difference of Values of Technology Components Between the Fi rst and Second Customer DHW12 DIN12 DHS12 DRP12 DHW12 1.000 (0.000) DIN12 0.357 1.000 (0.033) (0.000) DHS12 0.393 0.940 1.000 (0.013) (0.000) (0.000) DRP12 0.339 0.947 0.888 1.000 (0.051) (0.000) (0.000) (0.000) Source: database ODISSEO Bartlett-Chi Square Statistic: 255.834 Prob= 0.000 Note: probabilities are in brackets Firms'haracteristics, Product And Market Strategies By means of cluster analysis I'our kinds of subcontracting relationships were identified, depending on the feasible combination of the f'our technology comfionents. Summary statistics for thc four clusters are shown in Table 4. Table 4 Summary Statistics for the Four Clusters of Firms Variable DF DF F-Ratio Prob Hardware 3 55 108.61 0.000 Written Information 3 55 15.22 0.000 Human Skill 3 55 61.85 0.000 Rules and Procedures ' 55 25.31 0.000 Source: database ODISSEO 58 The four clusters show four diverse combinations of the components of technology which indicate four types of technological portfolios (Table 5). Table 5 Technological Ponfolfo and Firms'haracierisacs Cluster I Cluster 2 Cluster 3 Cluster 4 Anova Totaloffums=59 n=31 n=15 n=7 n=6 P HW =.03 HW =.22 HW =.05 HW =.35 IN=.35 IN=.28 IN=.38 IN=.25 HS =.29 HS =.19 HS =.07 HS =.07 RP = .33 RP = .30 RP = .51 RP = .33 AGE (years) 31.00 19.40 27.33 17.33 0.013 SIZE (employees) 255.42 43.5 23.0 29.3 0.489 SALES (fbiliion) 24.5& 3.67 2.68 2.187 0.065 INS ALES (%) 0.16 0.21 0.17 0.36 0.185 SALEMP 0.128 0.080 0.104 0.072 0.001 %PR I 0.63 0.67 0.79 0.5 0.008 %PR 123 0.92 0.97 0.92 0.89 0.413 REGMARK 0.53 0.75 0.74 0.87 0.006 ITAMARK 0.87 0.96 0.96 0.88 0.243 %CU12 0.69 0.76 0.57 0.83 0.034 FTECHI 1.712 2.408 1.134 1.522 0.001 FTECH2 1.270 1.805 0.652 1.061 0.041 DFTECH12 1.571 2.278 1.044 1.412 0.002 DCOMP12 0.142 0.096 0.98 0.019 0.05 DCOMP1234 0.370 0.480 0.222 0.648 0.135 Sowcei database ODISSEO In the first cluster, hardware embodied technology is not relevant. On the contrary, technology embodied as information and rules and procedures has a higher weight and technology tmnsferred thmugh people is much more important than in the other clusters. Most firms belonging to this cluster are main contractors in the automotive industry. These firms are involved in the development of a new model from the early stage in co-design with the customer. In this stage there are still problems that must be solved and uncertainty is high. Consequently, establishing an intense communication between customer and subcontractor design teams becomes a strong imperative for an effective development process. Three firms subcontract to the electronic industry. The second cluster presents a more balanced exchange technological portfolio, even though the rules/procedures component has a slightly higher weight relative to the others. Some primary firms subcontractors to the aerospac'e/defense industry are in this cluster. Two firms are second- tier subcontractors for the aerospace industry. One is a second tier subcontractor to the automotive industry. 59 The third cluster contains firms having an unbalanced exchange technological portfolio in which information and rules/procedures prevail. Generally, firms lx:longing to this cluster are subconuactors to traditional low-medium technology industrial sectors (elecuotechnology, texule-machinery building). These firms do not establish stable and constant relationships with their customers. Consequently, when they begin to collaborate with a new customer there are no incentives for this to transfer a relevant amount of technology to the subcontractor factory, such as practices for quality management, costs evaluation methods, and scheduling plans. Common standards and routines cannot be established and a large amount of codified technology must be transferred in every transaction. That explains the relative high value for information and rules. The last cluster is comprised of small firms subcontracting almost exclusively to the aerospace and defense industry, Technology embodied in raw material and specific equipment is relevant in this cluster. These fums receive pre-machined parts to manufacture fixtures, jigs, and gauges from the customer. A high level of standardization supports the transaction making it effective. Firms belonging to the first cluster are, on average, older and larger. Thirty-one years is the average age (AGE) for these firms. On average, their size (SIZE) is 255 employees and 24.58 billion of lire as sales per year (SALES) (no statistical evidence supfiorts this). As an indicator of success, the average firm's growth in sales over the last three years was chosen GNSALES). The values of this variable show that the small firms of the fourth cluster in 1990-91 were faster growing than the other firms. The positive demand uend of the aerospace industry in that period probably produced a "driving effect" on them. The investigated firms primarily serve Italian markets (ITAMARK), which represent almost the exclusive markets for firms in the second and third clusters (96 percent). Firms of the fourth cluster have plants that are localized close to their customers. Thus, their regional market (REGMARK) coincides with the Italian market. At the same time, they have international customers as firms of the first cluster. The internationalization process of the fums belonging to these two clusters has a different origin. The belonging to international groups favored the entrance into foreign markets of the automotive component subcontractors of the first cluster, which established collaborauve relationships with the competitors of their main Italian customer. Firms of the fourth cluster entered foreign markets primarily thanks to the international collaborative agreements engaged by their Italian customers. In lxith cases, this process of internationalization supported a circulation of technology in thc subeonuacting chains. A relatively high dependence on a few major customers and a certain product specialization characterize these firms. Firms of the third cluster present a lower dependence on the first customer (OUI) than other firms. They derive less than 40 percent of their sales from their first customer and less than 60 percent from their first two customers (%CU12). These firms had short- term contractual agreements with their customers. These results support that a high importance given to procedures during the transaction is a consequence of a low dependence on the customer and a scarce iniual fit also, an idea that emerged from the analysis of the exchange technological portfolio. On the contrary, these firms follow a strategy of product specialization. The 79 percent of the total sales is derived from only one line of products (%PRI). There are no substantial differences bctwecn firms of the first and second clusters as regards dependence on the customer. Firms of the fourth cluster are highly dependent on the first two customers, which sell in the same industry, but on the conuary, are less product specialized. These firms are small family-owned machine shops. Indeed, 29 employees and 2,187 billions of lire represent the average size. The low ratio of sales to employees (SALEMP) can be explained by the fact that 60 these firms receive from the customer the raw materials, or pre-machined parts, the cost of which is not included in the sales evaluation. These firms largely invested in numerically controlled equipment (usuafly, single NC machines or centers) in the last three years. The flexibility of the new technology made it possible for these firms to diversify production gaining efficiency. During the interview, respondents were asked to rate the technological sophistication of the product manufactured for the first three main customers on a 1-9 scale. To assess product technological sophistication, the following dimensions were considered: the level of conformity to quality specs requested by the customer for that product; the level of difficulty in realizing it; and the time required to leam how to do that product with the required degree of quality. To limit the effects of the subjectivity of judgments, the analysis was done considering the difference between the values of technological sophistication of products manufactured for the two main customers. This value was normalized by dividing it for the maximum score rate given by the respondent (DCOMP12). From the analysis, it emerges that there is no signiTicant difference between the technological sophistication of the product realized for the first two customers in the case of firms belonging to the first, second, and fourth clusters. Differences become higher passing from the product realized for the first two customers to products realized for the other customers. That is particularly the case of firms in the fourth cluster. These firms have the first two customers in the aerospace indusuy, while the others are in low technology industries. Some strategies of technological diversiTication can be thus identified. Firms in the second and fourth clusters pursue almost the same strategy. Indeed, most of the firms of the second cluster are subcontractors of the aerospace industry while all the firms of the fourth cluster subcontract to the aerospace industry. Technology flow largely increases with product sophistication as shown in Table 6. It appears that manufacturing more sophisticated products requires a deeper integration with the customer. Table 6 RelationBetweenTechnologyFlowandproductTechnologicalSophistication: FirstandSecond Customer. Dependent variable: DFTECH12 N = 56 Squared R = 0.443; no constant ~v' ~ffi i n ~rr r RDgtahiili (2 tail) DCOMP12 2.637 0.670 0.004 Sowcet database ODISSEO Note: In the sample 56 firms have more than one customer. 61 In the four clusters, the technological flow from the first customer (FTECHI) is considered more important than the flow from the second customer (FFECH2) in the acquisition of technological capabilities. The calculated values for the difference (DFTECH12) confirm this tendency. It is interesting to note how the values for both the flows are lower in the case of firms belonging to the third cluster. CONCLUSIONS This study is exploratory. Further invesugation is required m assess the validity of the considerations done. Several caveats should be identified before dmwing any conclusion. First, findings rely on subjecuve judgments and key variables are built from these judgments. A major ef fort should be undertaken to use a more objective scale to measure variables and more effective and objective methods to assign auribute weights. Second, a more stratified and larger sample should be chosen. The number of firms subcontracting to the aerospace and automotive industry is high relauve to the firms subcontract- ing to other industrial sectors. Additionally, cross-cultural studies are necessary to determine if the results can be generalized to other countries. Finally, a more reliable and sophisticated statistical analysis should be conducted. Cluster analysis remains an exploratory tool. Nevertheless, the study provides the basis for more refined research that can enable researchers to provide useful guidelines for subcontracting and customer firm's technicians and managers to better model their collaborative relationships. To summarize, findings presented in this paper suggest these main conclusions. The customer firm, in different ways, enables its subconuactors to achieve a higher technological capability. The constant request of better manufacturing performances and more sophisticated parts supplied, together with an intense technological flow, contribute to the upgrading of the subcontract's technological asset. The need to be more integrated with the customers, to increase efficiency, and particularly to increase product reliability and quality forced these firms to largely implement new production technology and to focus on new manufacturing methods. The deeper is the integration between the customer and the subcontractor, the more intense is the technological flow between the parties. This finding seems to confirm what emerged in other studies. A higher involvem& nt in the process of product development and manufacturing determines, but at the same ume needs, a more intense and articulated flow of technology between the parties. In some industries, customer/subcontractor uunsactions assume particular characterisucs and the technological content of the arrangement depends on the characteristics of the subcontracting firms. 62 REFERENCES Allen, TJ. (1977).Managing the Flow of Technology (1986).Cambridge, MA: the MIT Press. Daft, R.L.&. Lengel, R.H. (1986).Organizational information requirements, media richness and structural design. Management Science, 32(5), 1383-1403. Esposito E., and Raffa, M.(1991, June), Supply in hi-tech industry: the role of the small businesses. Proceedings of the 36th ICSB World Conference, Vienna. Galbraith, J.R. (1973).Designing Complex Organizations. Reading, MA: Addison Wesley. Halldn L., Johanson, J., gi Seyed-Mohamed, N. (1991,April). Interfirm adaptation in business relauonshi ps. Journal of Marketing, 55, (29-37. Harker P.T. & Vargas, L.G. (1987). The theory of ratio scale estimation: Saaty's analytic hierarchy process. Management Science, 33(11),554-571. Hayes, R H. & Wheelwright, S.C.(1984). Restoring our compeii ii ve edge: Competing Through Manufacturing. New York: Wiley. Lazerson, M. (1990). Subcontracting as an alternative organ isational form to vertical-integrated production. Depanment of Sociology, State University of New York. Nelson, R.R. & Winter, S.G.(1982).An Evolutionary Theory of Economic Change. Cambridge, MA: Harvard University Press. Saaty, T.L. (1980). The Analytic Hierachy Process. New York: McGraw Hill. Saget, E (1988). Partnership between small and large firms: the case of Japan. Partnership Between Small and large Firms. London: Grahm&Trotman. Starbuck, W.H. (1992). Learning by knowledge intensive firms. Journal of Management Studies, 29(6), 713-740. The Technology Atlas Team (1987), Components of technology for resources transformation. Technological Forecasting and Socioi Change, 32(1). Thompson J. (1967). Organizationsin Action. New York: McGraw Hill. Weick K.E. (1979). The Social Psychology of Organizing. Reading, MA: Addison Wesley. 63 APPENDIX Channels to Technology Components Contribution Map Raw material HW Pre-machined parts Equipment Training Advice on specific issues IN Collaboration at the launch of lhe order Written documents HS Meetings In-progress check Support for the RP adoption of QC procedures QC procedures Scale used to assess contribution weights: 1: no contribution at all 100: highest contribution 64