Journal of the Operational Research Society (1997) 48, 1131±1143

#1997 Operational Research Society Ltd. All rights reserved. 0160-5682/97 $12.00

Factors affecting international product design GC Hadjinicola1 and KR Kumar2 1

University of Cyprus and 2 University of Southern California

In this paper we examine the impact of changes of such factors as tariffs/import cost, exchange rate, and unit savings derived from economies of scale, on the product design of four international strategies which are characterised by two dimensions. The ®rst dimension describes whether the company offers a standardised or a customised product. The second indicates whether the company centralises its production to a single facility in one country or decentralises its production to facilities located in each country. To address the above issue, we present a model that has elements from marketing and manufacturing. For the case where the product has one attribute, we show that when tariffs/import cost decrease, an international enterprise should respond by enhancing the features of its products. Similarly, the product features should be enhanced when the exchange rate increases or the unit savings derived from economies of scale increases. Numerical examples indicate that an international enterprise should change its production con®guration from decentralised to centralised, in environments of high tariffs/import cost. Furthermore, an international enterprise should change its product policies from customised to standardised when the savings derived from economies of scale are high, and the exchange rate increases. Keywords: exchange rate; manufacturing policy; marketing policy; tariffs

Introduction In many industries, the battleground in global competition is shifting to rapid new product development, production and distribution. For example, in the personal computer industry, product lifecycles are less than 2±3 years. Within this time frame, product design/launch, production rampup/global distribution, and production ramp-down/inventory mop-up have to occur ¯awlessly. A one-month delay in product design critically affects the product launch, `eating' into the high-pro®t phase of the life cycle, making the difference between a ®nancial success and failure. One of the critical success factors in such competition is creating a tight integration between product design and production con®guration strategies that can react quickly to changes in the market environment. Product design strategies are strongly affected by advances in technology. Speci®cally, improvements in communication, transport, and travel are driving consumers around the world to be aware of, and demand the same set of products.1 Facing this phenomenon termed globalisation of markets, Levitt states that `companies must learn to operate as if the world were one large marketÐignoring super®cial regional and national differences.' Nevertheless, the belief of the existence of homogeneous markets and therefore, the offering of a standard product design worldwide, has been challenged by the traditional paradigm that Correspondence: Dr GC Hadjinicola, Department of Public and Business Administration, School of Economics and Management, University of Cyprus, Kallipoleos 75, PO Box 537, CY 1678, Nicosia, Cyprus. E-mail: [email protected]

consumers across countries have different needs and therefore, demand different sets of products.2,3 Two key global product design strategies, that have been researched extensively in the literature, are standardisation and customisation. The debate whether to provide consumers around the world with the same standardised product or with a custom-tailored product originated in the early sixties.4 So far, neither viewpoint has managed to dominate since both of them have examples of successes and failures as described by Levitt1 and Kashani.5 The primary bene®ts derived from product standardisation are savings in cost due to high production volumes. Low production costs ensure lower prices but at the risk that the standardised product may not meet the needs of consumers in a speci®c market. On the other hand, product customisation provides consumers with a product of their preference but perhaps at higher prices due to lower volume. Intricately intertwined with the notions of product standardisation/customisation is the con®guration and coordination of the production function in an international setting. Con®guration deals with plant location issues across countries and coordination refers to the alignment of the geographically dispersed activities.6 Production and supply chain con®guration are key activities that global corporations are undertaking today. As such, product standardisation could be facilitated by centralised production, whereas product customisation could be promoted through localised production. Three factors that have started to affect global manufacturing strategies involving production con®guration and coordination are tariffs/import cost, exchange rates, and economies of scale. Changes in tariffs and government

1132 Journal of the Operational Research Society Vol. 48, No. 11

subsidies are having a major impact on manufacturing strategies. The 1994 GATT treaty has tried to uniformly reduce tariffs and government subsidies; the reduction of tariffs in various countries offers the opportunity to centralise plants and economise. Similarly, the recent US±Japan trade controversy on automobile parts threatened to impact production and purchase strategies in the Japanese automobile industry, by making production in the US more attractive. In recent times, exchange rates have played a signi®cant role in the production con®guration issue. For example, one of the primary motivations for BMW to disperse its production plants around the world is the strong valuation of the mark, relative to the US dollar. Since almost all its production costs are in marks and its biggest source of revenues is the US market, BMW decided to open a plant in Spartanburg, South Carolina. The organised pursuit, retention, and storage of knowledge has become the next challenge for businesses, continuing on the coat-tails of the quality revolution. One of the clear bene®cial outcomes of this pursuit could be the strengthening of economies of scale and scope. There has been considerable research on the impact of economies of scale on the production con®guration issue.7 High volume production, at a centralised location, leads to a lower unit cost of production and retained knowledge drives it even lower. Dispersed production leads to problems of transfer of technology/knowledge across various units in the organisation. While it is clear that product design issues and economic factors affect production strategies, what is the impact of these economic factors on product design? Besides the impact of exchange rates on facility location, should International Enterprises (IE) change their product design as the exchange rate ¯uctuates? Should changes in tariffs/import cost that affect global production con®guration also impact global product designs? Should changes in savings derived from economies of scale that lead to facility rationalisation8 also force product redesigns? Even though manufacturing and product design should not be impacted by short term exchange rate and tariff changes since these changes occur faster than the speed at which ®rms can build plants in new countries or design new products, understanding the impact of ¯uctuations and changes across these variables can be useful if these changes can be forecasted. As such, an IE in formulating its long term manufacturing and product design strategy should take into consideration the forecasted changes in the above variables. In this paper, our attempt is to address the above questions through four international strategies characterised by two dimensions associated with the location of the production facilities and the product design that an IE can adopt. The ®rst dimension describes whether the company offers a standardised or a customised product

design. The second dimension indicates whether the company centralises its production to a single facility in one country or decentralises its production to facilities located in each country. Figure 1 describes the above two dimensional framework and provides the abbreviated names for each strategy. To address the above questions, we present a model that has elements both from marketing and manufacturing. Speci®cally, the model incorporates such marketing concepts as product positioning, consumers' ideal point, and product attraction, as well as production concepts such as savings derived from economies of scale, and centralised/decentralised production. The model also incorporates such variables as the exchange rate between two countries and tariffs/import cost. For the sake of exposition and simplicity, the modelling framework presented deals with a two country scenario. For the case where the product has one attribute dimension, we show that international strategies affected by changes in tariffs/import cost, exhibit an increase in the product's attribute upon a decrease in tariffs/import cost. In addition, we show that increases in the exchange rate results in an increase in the product's attribute. Further, increases in the unit savings derived from economies of scale result in an increase in the product's attribute. A numerical example shows that these increases exhibit a levelling effect where additional bene®ts from the above phenomena do not translate into returns from additional product features. The numerical examples also indicate that an international enterprise should change its production con®guration from decentralised to centralised in environments of high tariffs/import cost. Furthermore, an international enterprise should change its product policies from customised to standardised, when the savings derived from economies of scale are high and the exchange rate increases. The organisation of the paper is as follows. First, we include the model formulation and then we present the formulation of the pro®t maximising programs of the four international strategies. The next section contain a theore-

Figure 1 Four international production-product strategies.

GC Hadjinicola and KR KumarÐFactors affecting international product design

tical analysis on the effect of tariffs/import cost, exchange rates, and economies of scale on product design. The last two sections include numerical examples, concluding remarks and directions for future research. Model formulation Positioning a product is usually aimed at offering a product which meets the needs of consumers in a particular market segment in order to increase the ®rm's market share and pro®tability. Factors in¯uencing the positioning of a product include consumer attraction derived from the product's position and price, cost of furnishing a product with its attributes, and production volume/economies of scale. To capture the interrelationship of the above factors, we utilise the notion of a product attribute space where products can be abstractly represented as a set of coordinates. Each dimension in the attribute space designates a product characteristic, for example, the level of sweetness of a chocolate which translates into the per unit volume sugar content. Shocker and Srinivasan9 state that attributes should be actionable, that is, `indicate speci®c actions the manufacturer must take to build the product.' Under this framework, each consumer is assumed to have a set of most preferred attributes termed the ideal point. The greater the proximity of the product in the attribute space to the consumer's ideal point, the greater the probability that the consumer will purchase the product. This type of probabilistic consumer buying behaviour is only one of many found in the literature of probabilistic and deterministic choice models.10 Attraction models Product attraction is a well established notion in the marketing literature. As the distance between consumers' ideal point and the coordinates of the product offered in the attribute space decreases, attraction, and the propensity of consumers to purchase the product, increases. Attraction has also been assumed to be a function of other factors such as price, advertising and distribution expenditures. For example, as the price of the product increases, attraction decreases. On the other hand, as advertising and distribution expenditures increase, attraction increases. As such, attraction can be viewed as a measure of the `willingness' of consumers to purchase a product. In Gruca et al.11 attraction has been de®ned as Attr ˆ f …x†g…P† in which f …x† is a function of the product position x (a vector containing the attribute coordinates of the product) relative to the market's ideal point x* (a vector containing the coordinates of the ideal point). In general, the smaller the distance, the higher the value of the function. g…P† is a downward-sloping function of price P. In this paper, we

1133

neglect the effects of advertising and distribution in order to simplify the analysis. Using a formulation similar to the one by Shocker and Srinivasan,9 who modeled the distance between the product and the ideal point as a weighted Euclidean distance, we de®ne the function fj …xj † for a subsidiary in country j as L P

fj …xj † ˆ 1 ÿ

iˆ1

wij …x*ij ÿ xij †2 L P iˆ1

wij x*ij 2

ˆ1ÿ

…x*j ÿ xj †0 Wj …x*j ÿ xj † ; x*j 0 Wj x*j …1†

where xj and x*j are the L-dimensional vectors containing the attribute coordinates of the product and ideal point in country j… j ˆ 1; 2†, respectively. Their elements are denoted by xij and x*ij …i ˆ 1; . . . ; L†, respectively. Wj is an L 6 L diagonal matrix whose elements …wij † are the weights associated with attribute i in country j. In the above formulation we assume that 0 4 xj 4 x*j which further implies that 0 4 fj …xj † 4 1. The reason for restricting the space of the vector containing the product attributes is that the model presented in this paper is valid only in that range. One way to justify such a restriction is the fact that products with attributes greater than the ideal point will have a limited appeal to consumers due to the super¯uity of their features and the higher price caused by higher production costs. This will make the products unpro®table to produce. In order to simplify the analysis, we assume that the market has only one ideal point. In the product positioning literature, the optimal product position is usually derived after considering ideal points for each and every consumer.12 In this study, we assume that consumers in each country are homogeneous with respect to their preferences, distributed around the country's ideal point. Assuming a unimodal distribution for consumer preferences is a limitation. Nevertheless, studies such as Sudharshan et al13 consider one or more ideal points for each market segment. Eliashberg and Manrai14 presented a model for a new product positioning in a single market, using a single ideal point being the median value of consumers' ideal points. We de®ne the function gj …Pj † ˆ …1 ÿ …bj Pj =Yj ††, where Pj is the price of the product, bj is the price elasticity, and Yj is the average consumer income in country j. The above function follows by linearising the nonlinear function g…P† ˆ …Y ÿ P†b used by Gruca et al.11 We assume that Pj 4 Yj =bj , where Yj =bj is country's j reservation price which further implies that 0 4 gj …Pj † 4 1. Therefore, the product attraction for a subsidiary operating in country j is de®ned as Attrj ˆ fj …xj †gj …Pj † " #" # …x*j ÿ xj †0 Wj …x*j ÿ xj † bj Pj 1ÿ : ˆ 1ÿ x*j 0 Wj x*j Yj

…2†

1134 Journal of the Operational Research Society Vol. 48, No. 11

Under the above two assumptions concerning the product's attribute levels and price, it follows that the product attraction assumes values in [0, 1]. As a result, we can think of the attraction as the probability that a consumer in country j will purchase the product offered by the subsidiary serving the country. Average cost From the above de®nitions, we can see that the position of the product in the attribute space affects the attraction of the product. We assume that attraction has an immediate impact on the number of units demanded. The quantity produced determines the potential savings derived from economies of scale which further affects the average cost. The cost of production in turn affects the pricing policy of the ®rm and, as a consequence, the overall attraction of the product. In addition, the cost of production depends on the nature of the product which is determined by the product's position in the joint attribute space. One may assume that the larger the coordinate along an attribute dimension, the higher the production cost. This assumption is based on the fact that furnishing a product with a larger quantity of a particular attribute requires a higher cost. Studies by Choi et al,15 and Bachem and Simon16 employ this assumption by using a cost function that increases linearly with increasing attribute levels. Following the discussion above, we de®ne the average cost for a subsidiary in country j producing a product with attribute vector xj as ACj ˆ b0j
xj ÿ Kj Attrj Qj ;

Production and product design strategies for international enterprises Strategy 1: decentralised production±customised product (DP±CP) This strategy is synonymous to a multinational corporation which follows a decentralised policy under which local subsidiaries are given the autonomy to devise their own product and pricing policies. Following the characteristics of a multinational corporation as described by Hout et al17 and Yip,18 we assume that the responses of the subsidiaries in their respective countries are independent and, in particular, their pro®t functions are independent. Because subsidiaries are strategically independent and operations are autonomous, production of the product in each local market is the sole responsibility of the subsidiary serving the market. The pro®t function of a subsidiary in country j of an IE adopting the DP±CP strategy is given by Pj ˆ ‰Pj ÿ …b0j
xj ÿ Kj Attrj Qj †ŠAttrj Qj ;

…4†

Since production in country 2 takes place locally, the exchange rate r (de®ned as the number of units in the currency of country 1 we would obtain from the conversion of one unit of the currency of country 2) affects the total pro®ts generated by the subsidiary in country 2. As a result, an IE adopting the DP±CP strategy will have the following pro®t maximisation program PDPÿCP ˆ ‰P1 ÿ …b01
x1 ÿ K1 Attr1 Q1 †ŠAttr1 Q1

max

P1 ;P2 ;x1 ;x2

‡ r‰P2 ÿ …b02
x2 ÿ K2 Attr2 Q2 †ŠAttr2 Q2 :

…3†

where bj is an L-dimensional vector containing the unit costs for furnishing a product with each attribute i in country j (its elements are denoted by bij ), Kj are the unit savings derived from economies of scale in country j, and Qj is the sales potential in country J. Note that Qj is ®xed and is not in¯uenced by the producer's actions, such as changing the product price. We assume that the cost for a subsidiary to furnish a product with its attributes (xj ) is given by the dot product b0j
xj . Thus, the enhanced product features require more effort to be furnished, resulting in higher production cost. Note that we assume a linear relationship between the cost of furnishing a product with its attributes xj , and the magnitude of these attributes but nonlinear formulations could also be used. The average cost function also includes the term Kj Attrj Qj which represents the savings obtained from producing a volume of Attrj Qj units. We assume that Attrj Qj represents the sales volume. In this paper, in order to simplify the analysis, we use the above linear relationship to represent the savings derived from economies of scale. Hayes and Wheelwright7 provide an excellent discussion of concepts related to economies of scale.

j ˆ 1; 2:

…5†

Due to the decentralised nature of the DP±CP strategy, shown in (5), the pro®t maximisation program of an IE adopting the DP±CP strategy degenerates to the pro®t maximisation programs of each subsidiary.

Strategy 2: centralised production±customised product (CP±CP) Under this strategy, the IE manufactures products which are customised to the preferences of consumers in the local markets. Production of both products takes place in a central facility located in country 1. The product targeted for country 2 is shipped from the facility in country 1. The managerial objective of an IE adopting the CP±CP strategy is the maximisation of total pro®ts derived from the subsidiaries in the two countries. This pro®t is given by max

P1 ;P2 ;x1 ;x2

PCPÿCP

ˆ ‰P1 ÿ …b01
x1 ÿ K1 Attr1 Q1 †ŠAttr1 Q1

‡ ‰rP2 ÿ …b01
x2 ÿ K1 Attr2 Q2 † ÿ cŠAttr2 Q2 :

…6†

GC Hadjinicola and KR KumarÐFactors affecting international product design

Note that under the CP±CP strategy, the exchange rate r affects only the revenues obtained from country 2 since production of the customised products takes place in country 1. In addition, the pro®t margin for country 2 is decreased by the tariffs/import cost c (which includes the transportation cost). Strategy 3: centralised production±standardised product (CP±SP) The IE following this strategy offers a standardised product in the two countries in which it operates, with production taking place in a centralised facility located in country 1. This strategy requires centralised decision-making since the IE needs to design and produce a product with such features that yield the maximum total pro®ts. This strategy follows a philosophy where the world is viewed as a single entity/ market characterised by the presence of worldwide coordination. The standardised product is denoted by x, an L  1 vector (its elements are denoted by xi ) containing the product's attribute coordinates. An IE adopting the CP±SP strategy will maximise the total pro®ts generated from both markets: max PCPÿSP ˆ ‰P1 ÿ …b01
x ÿ K1 …Attr 1 Q1 ‡ Attr2 Q2 ††Š

P1 ;P2 ;x

 Attr1 Q1 ‡ ‰rP2 ÿ …b01
x ÿ K1 …Attr1 Q1 ‡ Attr2 Q2 †† ÿ cŠAttr2 Q2 : …7† Since the standardised product is centrally produced in country 1 and then shipped to country 2, the average cost will be a function of the total volume produced for both countries. Since the product is manufactured in country 1, it follows that b1 and K1 will determine the average cost of the standardised product. The average cost for country 2 is also burdened by tariffs/import cost. In addition, the revenues obtained from country 2 are affected by the exchange rate. Strategy 4: decentralised production±standardised product (DP±SP) Under this strategy, the IE offers a standardised product to the consumers of the two countries. The standardised product is manufactured at production facilities located in both countries. As a result, each production facility offers the local market it serves the standardised product. An IE adopting the DP±SP strategy will have the following pro®t maximisation program max PDPÿSP ˆ ‰P1 ÿ …b01
x ÿ K1 Attr1 Q1 †ŠAttr1 Q1

P1 ;P2 ;x

‡ r‰P2 ÿ …b02
x ÿ K2 Attr2 Q2 †ŠAttr2 Q2 : …8†

1135

From (8) we see that the pro®ts obtained in country 2 are affected by the exchange rate.

Factors affecting international product design The pro®t maximisation programs of the four strategies presented in (5), (6), (7), and (8) are nonlinear and no closed form solution for prices and product positions can be derived. This is a common drawback of all product positioning formulations that incorporate the nonlinear distance of the product's position to the ideal point. Hadjinicola and Kumar19 present conditions that the above four pro®ts functions must meet in order to preserve concavity of the pro®t functions in terms of prices and product positions, and thus ensure the existence of an optimal solution. In the propositions to follow, we simply assume concavity. In particular, @2 P=@x2ij < 0, a necessary but not suf®cient condition for concavity. Furthermore, we shall use the fact, following from (2), that for 0 4 xj 4 x*; j @Attrj =@xij > 0. The following proposition characterises the pro®tability relationship between the DP±CP and the DP±SP strategies. Proposition 1 The pro®ts generated by an IE adopting the decentralised production and customised product (DP± CP) strategy are higher or equal to the pro®ts generated by the decentralised production and standardised product (DP±SP) strategy. Proof

See Appendix.

Since the DP±CP and DP±SP strategies follow decentralised manufacturing and thus both face the same production parameters in each country, the differentiating factor affecting the IE pro®tability is the product offered by each strategy. Under the DP±CP strategy, the IE manufacturers in each country a customised product that closely meets the needs of the local consumers. On the other hand, an IE adopting the DP±SP strategy manufactures and offers in each country a standardised product. This product may not necessarily meet the needs of local consumers. Furthermore, decentralised production and therefore, lower volumes, does not allow the DP±SP strategy to derive signi®cant bene®ts from economies of scale as the CP±SP strategy. As a result, the DP±SP strategy generates equal or less pro®ts than the DP±CP strategy. The above comparison of the two strategies does not take into consideration the costs of designing the products. For example, strategies offering a standardised product may bene®t from economies of design, thus allowing for lower product prices that would lead to higher demand and perhaps higher pro®ts. On the other hand, strategies offering customised products will be burdened by the cost of designing more products and the amortication of such cost

1136 Journal of the Operational Research Society Vol. 48, No. 11

will yield higher product prices, lower demand and maybe lower pro®ts. The propositions that follow examine the sensitivity of the optimal product designs of the three strategies left to consider (since DP±SP is dominated by DP±CP) upon changes in tariffs/import cost, exchange rate, and unit savings derived from economies of scale. We start our analysis with the simplest case, based on two assumptions. First, the product has only one attribute dimension and second, prices are ®xed. Subsequently, we provide a general approach to examine the impact of parameter changes on both prices and attribute levels for products with more than one attribute. Additional discussion of the propositions is presented in the numerical section. Proposition 2 Given prices, when tariffs/import cost increase (decrease), strategies affected by the change will exhibit a decrease (increase) in the product's attribute. Proof

See Appendix.

More speci®cally, when tariffs/import costs increase, the products' attribute in countries 1 and 2 of the DP±CP strategy remain unchanged since no inter-country transfers are involved. The product attribute of the CP±CP strategy remain unchanged in country 1 and decrease in country 2. Finally, upon an increase in tariffs/import cost, the attribute of the standardised product of the CP±SP strategy decreases in both countries. In order to increase the attraction of its product in the country affected by the increase in tariffs/ import cost, the IE is forced to decrease the features of the product in order to decrease its production costs and maintain its pro®tability. Proposition 3 Given prices, when the exchange rate between the two countries increases (decreases), strategies affected by the change will exhibit an increase (decrease) in the product's attribute. Proof

See Appendix.

Speci®cally, when the exchange rate between the two countries increases, the product attribute of the DP±CP strategy in countries 1 and 2 is not affected. The product attribute of the CP±CP strategy remains unchanged in country 1 and increases in country 2. Finally, for the CP± SP strategy, an increase in the exchange rate between the two countries will result in an increase of the attribute of the standardised product in both countries. These changes follow from a similar rationale as in the tariffs/import cost case. Proposition 4 Given prices, when the unit savings derived from economies of scale increases (decreases), all three strategies will exhibit an increase (decrease) in the product's attribute in both countries. Proof

See Appendix.

Proposition 4 indicates that the additional savings derived from economies of scale are passed on to the consumers by the enhancement of the products with additional features. The additional cost of increasing the products' features is balanced by savings derived from economies of scale, leaving the strategies with a more attractive product. For cases where the product has more than one attribute dimensions, the assessment of the impact of parameter changes on both prices and product attributes needs to follow a different approach. Speci®cally, for each international strategy described in this paper, de®ne @P=@xij ˆ qij and @P=@Pj ˆ pj for i ˆ 1; . . . ; L; j ˆ 1; 2. These functions can be derived from the pro®t functions of each international strategy. For strategies adopting a standardised product across the two countries we de®ne @P=@xi ˆ qi . The ®rst order conditions imply that qij ˆ 0 or qi ˆ 0 and pj ˆ 0. Using the chain rule, total differentiation of the above functions with respect to one of the parameters, for example y, gives the following system of equations, presented in matrix form: 2 @p1 =@P1 @p1 =@P2 @p1 =@x11


@p1 =@xL1 6 @p2 =@P1 @p2 =@P2 @p2 =@x11


@p2 =@xL1 6 6 @q11 @P1 @q11 =@P2 @q11 =@x11


@q11 =@xL1 6 6 .. .. .. .. 6 . . .


. 6 6 @qL1 =@P1 @qL1 =@P2 @qL1 =@xL1


@qL1 =@xL1 6 6 @q12 =@P1 @q12 =@P2 @q12 =@x11


@q12 =@xL1 6 6 .. .. .. .. 4 .


. . . @qL2 =@P1

@qL2 =@P2

@p1 =@x12 @p2 =@x12 @q11 =@x12 .. .






@qL1 =@x12






@q12 =@x12 .. . @qL2 =@x12

















@qL2 =@xL1




@qL2 =@xL1

2 3 3 @P1 =@y @p1 =@y 6 6 @p =@y 7 7 @p2 =@xL2 7 6 2 76 @P2 =@y 7 7 6 76 7 7 6 @q11 =@y 7 @q11 =@xL2 76 @x11 =@y 7 6 76 7 7 6 76 7 7 .. .. 6 6 7 7 7 . . 6 76


7 7 76 7 ˆ ÿ6 7: 6 @qL1 =@y 7 @qL1 =@xL2 76 @xL1 =@y 7 6 76 7 7 6 6 @q =@y 7 7 @q12 =@xL2 7 6 12 76 @x12 =@y 7 7 6 76 7 7 .. .. .. 6 76 7 7 4 54 5 5 . . . @p1 =@xL2







@qL2 =@xL2

32

@xL2 =@y

@qL2 =@y

The above system of equations can be written as Ax ˆ ÿb. The vector x contains the directions of change at the optimal pro®t point of all variables such as prices and product attributes with respect to some parameters y. This parameter can be c, r or Kj . The direction of change of the variables with respect to the parameter can not be stated explicitly, as was done in the analysis where the product has one attribute, because of the complexity of the functions and dimensions of the problem. Numerical techniques can assist us in observing the direction of change of the variables with respect to the parameters.

GC Hadjinicola and KR KumarÐFactors affecting international product design

Numerical example The dif®culty in obtaining closed form solutions of the optimal prices and products of the three dominant international strategies points to the use of numerical methods. In this paper, we employ Newton's successive relaxation method (Schwartz20, p. 224) to compute the optimal prices and product attribute levels. We provide a numerical example to examine the impact of changes in tariffs/import cost, exchange rate, and unit savings derived from economies of scale on the product offered by the three interesting strategies. The numerical example also serves to examine the impact of parameter changes on the pro®tability of the strategies. The numerical example is carried out in the form of a sensitivity analysis where all parameters are ®xed at a speci®c value, except one whose value is allowed to vary. The numerical example involves the simple case where the product has a single product attribute dimension. If not otherwise stated, the parameters take the following values:

c ˆ 5; w2 ˆ 1;

r ˆ 1; Y1 ˆ 100; Y2 ˆ 200; w1 ˆ 1; Q1 ˆ 200; Q2 ˆ 300; K1 ˆ 0:03;

K2 ˆ 0:03; x* 1 ˆ 5;

b1 ˆ 1:6;

x* 2 ˆ 6:

Figure 2

b2 ˆ 3;

b11 ˆ 2;

b12 ˆ 2;

1137

Figure 2 presents the attribute levels of the products offered in both countries by all three strategies when tariffs/ import cost increase. From the ®gure we note that in country 1, the product design of strategies using the customised product policy is not affected by changes in tariffs/import cost. This is due to the fact that production takes place in country 1 and the product does not have to be transported. For country 1, we see that the product features of the CP±SP strategy decrease when tariffs/import cost increase. This phenomenon can be attributed to cost offsetting. The increase in cost due to the increase in tariffs/ import cost forces this IE to offer a standardised product with lower product features in order to reduce the cost of production. Similar phenomena occur for the CP±SP and CP±CP strategies in country 2. Centralised production adopted by both strategies obviously is affected by changes in tariffs/ import cost. Figure 2 shows that the rate of decrease in product features, when tariffs/import cost increase, is higher for the CP±CP strategy than the CP±SP strategy. This is due to the fact that the CP±CP strategy lacks the savings derived from economies of scale that the CP±SP strategy enjoys. The latter strategy offers a standarised product and is less sensitive to the increase in tariffs/ import cost due to the high production volumes. Figure 3 presents the changes in product features for both countries when the exchange rate between the two coun-

The product features in the two countries for three international strategies with varying tariffs/import cost (c).

1138 Journal of the Operational Research Society Vol. 48, No. 11

tries increases. In country 1, the product features of strategies adopting customised product policies are not affected by the exchange rate since production takes place locally in country 1. In country 1, the product features of the CP±SP strategy increase when the exchange rate increases. The higher value of the currency in country 2 increases the proportion of the pro®ts derived from country 2. This forces this IE to offer a standardised product with more features in order to meet the product needs of country 2s consumers. In country 2, an increase in the exchange rate does not affect the product offered under the DP±CP strategy since production takes place in country 2 as well. Decentralised production causes the pro®t maximisation program of an IE adopting the DP±CP strategy to degenerate to the pro®t maximisation programs of each subsidiary. On the other hand, an increase in the exchange rate causes an increase in product features for the two strategies adopting centralised production. The reason for such an increase in product features is the same given for the CP±SP strategy in country 1 above. The issue here is the rate at which the product features increase. For the chosen set of parameters, the CP± CP strategy has the highest rate of increase. This occurs since this IE tries to customize the product for country 2 as much as possible due to its signi®cant dependence on country 2 for its pro®tability. Figure 3 also shows a

levelling effect representing the point at which there are no returns to additional increases in the product features due to increases in the exchange rate. Figure 4 shows changes in product features when the unit savings derived from economies of scale varies. Speci®cally, an increase in the unit savings derived from economies of scale causes an increase in the total savings derived from economies of scale. These savings are passed on to the consumer in the form of a more enhanced product, priced at such a level to increase the IEs overall pro®tability. Figure 4 also shows a levelling of the product features when the unit savings derived from economies of scale increases. The levelling represents the point where there are no returns on additional increases in the features of the product. This is caused by the offsetting of the bene®ts derived from the increased product features, by higher prices resulting from higher production costs. The point at which the levelling begins depends on other parameters such as the value of tariffs/import cost. Figure 5 presents the optimal pro®ts generated by the three strategies when tariffs/import cost, exchange rate, and unit savings derived from economies of scale change. The range of the three parameters (c; r; Kj ) is the same as the one used in the sensitivity analysis above. Even though the results of the sensitivity analysis are dependent on the

Figure 3 The product features in the two countries for three international strategies with varying exchange rate (r) between the two countries.

GC Hadjinicola and KR KumarÐFactors affecting international product design

1139

Figure 4 The product features in the two countries for three international strategies with varying unit savings derived from economies of scale (K1 ˆ K2 ).

choice of the parameter set, Figure 5 provides certain managerial insights. Primarily, it is evident that none of the strategies is uniformly dominant when tariffs/import cost vary. The ®rst graph in Figure 5 suggests that in environments of high tariffs/import cost, strategies adopting decentralised production generate higher pro®ts than strategies adopting centralised production. The reason is that in strategies with centralised production, the shipment of the product, whether customised or standardised, from the country it is manufactured to the other country, causes an increase in its price. The lower demand and the reduction of the pro®t margin in the second country, due to the introduction of tariffs/import cost, is re¯ected in lower pro®ts. Figure 5 also shows that the total optimal pro®ts generated by an IE adopting decentralised production are not affected by tariffs/ import cost since production takes place locally. Of particular interest is the point at which the pro®t curves of the strategies adopting decentralised and centralised production schemes intersect. This is the point at which it becomes more pro®table for the IE to change its production strategy. For example, in environments of low tariffs/import cost it is more pro®table to centralise production and exploit the bene®ts from economies of scale. The analysis presented in this paper does not take into consideration the investment required to locate an additional production facility in the second country for strategies

adopting decentralised production. If we account for the cost of setting up a new facility, the result will be a movement to the right of the point at which the pro®t curves of the various strategies currently intersect, namely to higher values of tariffs/import cost. The second graph of Figure 5 presents the pro®ts of all strategies when the exchange rate of the currency between the two countries varies. The ®gure shows that as the exchange rate increases the pro®ts generated by all three strategies increase. Consider for example the CP±SP strategy. The increase in pro®ts comes from two sources. First, pro®ts increase due to the increased value of the currency in country 2. Second, since the contribution of country 2 to this IEs total pro®ts increases, the IE is forced to place more emphasis on the needs of the consumers in country 2. The result is a standarised product with increased product features. The enhanced product meets the needs of consumers in country 2 but, at the same time, the higher volume produced assists in the reduction of prices. This also bene®ts consumers in country 1 and pro®tability increases in country 1 as well. This phenomenon is particularly evident at high values of the exchange rate where the CP±SP strategy generates higher pro®ts than the other two strategies. Here, the effect of the exchange rate dominates the effect of tariffs/import cost. At lower values of the exchange rate, the effect of

1140 Journal of the Operational Research Society Vol. 48, No. 11

Figure 5 Total pro®ts of three production-product strategies with varying tariffs/import cost, exchange rate, and unit savings derived from economies of scale (K1 ˆ K2 ).

other factors such as tariffs/import cost dominates, thus allowing strategies adopting decentralised production con®gurations to generate higher pro®ts. The points at which the pro®t curves intersect indicates the need for a change in the production structure and product policies of the IEs. For strategies adopting customised product policies (DP±CP, CP±CP), the increase in pro®ts caused by increases in the exchange rate results only from changes in the pro®ts generated from country 2. This is also evident from the pro®t functions of the two strategies presented in the previous section. For example, since production in the

CP±CP strategy takes place in country 1, the increase in pro®ts from country 2 is caused by the revenues obtained from country 2 when translated into the currency of country 1. Finally, the third graph of Figure 5 contains the total pro®ts generated by the three strategies when the unit savings derived from economies of scale vary. In this ®gure we assume that the unit savings in both countries are the same …K1 ˆ K2 †. The ®gure shows that as the unit savings derived from economies of scale increase, total pro®ts generated by an IE under any of the three strategies

GC Hadjinicola and KR KumarÐFactors affecting international product design

increase. At high values of unit savings derived from economies of scale, the centralised production and standardised product strategy does not only generate higher pro®ts but also has a higher rate of pro®t increase than the other two strategies. The standardised product is the deciding factor since the higher volume produced by the centralised facility enables the IE to signi®cantly reduce cost through high volumes. On the other hand, in environments of low savings derived from economies of scale, the deciding factor determining pro®tability becomes the nature of the product design and the value of tariffs/import cost. The bene®ts realised from centrally producing a standardised product are outperformed by the offering of a customised product which is locally produced. This bene®ts the IE in two ways. First, the offering of a customised product that meets the needs of local consumers and second, the reduction of the overall cost by the removal of tariffs/import cost through local production. As a result, in environments of low savings derived from economies of scale, the DP±CP strategy outperforms the CP±SP strategy. In such an environment, the CP±CP strategy appears to generate lower pro®ts than the DP±CP strategy. Since the latter two strategies offer customised products the deciding factor for pro®tability is tariffs/ import cost (that burdens the products of the CP±CP strategy). Conclusions Two critical issues in international manufacturing are product design and the con®guration/coordination of production activities. Speci®cally, the design of standardised or customised products is a central issue for companies competing internationally. Offering a custom-tailored product to each country has the advantage of meeting the needs of local consumers. The disadvantage of product customisation is the lack of high production volumes which do not enable an international enterprise to fully exploit economies of scale. On the other hand, product standardisation allows an international enterprise to bene®t from a high worldwide volume, but at the risk that in some countries the standardised product might not meet the needs of local consumers. The location and coordination of the production facilities is also a crucial issue for international enterprises. This relates to the extend to which international enterprises centralise or decentralise their production in a single or multiple facilities, respectively. In this paper, our attempt has been to see how changes in tariffs/import cost, exchange rate ¯uctuations, and changes in the unit savings derived from economies of scale affect the product design of international enterprises. To address the issue, we present a model that has elements from both marketing and manufacturing. We examine four strategies that international enterprises can adopt concerning the location of the production function and the product

1141

design. These strategies are: decentralised production-standardised product (DP±SP), centralised production-customised product (CP±CP), centralised production± standardised product (CP±SP), and decentralised production-standardised product (DP±SP). We show that the DP±CP strategy is superior to DP±SP in terms of pro®ts. More importantly, for the simple case where the product has one attribute dimension, we show that strategies affected by changes in tariffs/import cost, exhibit an increase in the product attribute upon a decrease in tariffs/import cost. In addition, we show that increases in the exchange rate results in an increase in the product attribute. Furthermore, an increase in the unit savings derived from economies of scale results in an increase in the product attribute. From the numerical example we observe that an international enterprise should change its production con®guration from decentralised to centralised in environments of high tariffs/import cost. Furthermore, an international enterprise changes its product policies from customised to standardised when the savings derived from economies of scale are high, and the exchange rate increases. This study is certainly not without limitations. Firstly, the assumption that consumer preferences in a country are homogeneous and are characterised by the market's ideal point is restrictive and does not consider the various market segments. Research should be extended to include distributions of preferences and thus non-homogeneous consumers. Secondly, the model is static since it does not incorporate time. For an international enterprise, time is important since the formulation of long term manufacturing and product design strategies should take into consideration the forecasted changes in the variables considered in this paper. Thirdly, even though the paper provides some analytical generalisable results, more work is required to substantiate the conclusions drawn from the numerical part of the paper. Given that international product design is affected by tariffs/import cost, exchange rate, and unit savings derived from economies of scale, an interesting direction of future research is the timing and frequency of product change when the above factors change. For example, given a forecast for the exchange rate, when it is optimal for a company to redesign its product? Given such an exchange rate forecast, when is it optimal to invest in new technology/training to improve savings from economies of scale, in order to offset the cost of redesign? Even though we show the impact of changes in tariffs/import cost, exchange rates, and unit savings derived from economies of scale on product design, the issue that remains to be examined empirically is whether companies are willing to reduce the features of their products when the changes in the parameters suggest so. In other words, product redesign might exhibit a `stickiness' where product upgrading is favored when required but product downgrading is not.

1142 Journal of the Operational Research Society Vol. 48, No. 11

Another direction for future research is to examine a hybrid approach of the polar extremes of complete standardisation/ customisation. A hybrid approach may exploit the best of both worlds by centrally producing a `core' product which can be adapted to better meet the needs of local consumers.21 AcknowledgementÐThe authors would like to thank the two anonymous referees whose comments have improved this paper.

Appendix

Using (11), (12), and the assumption that @2 P=@x21j < 0 and that @Attrj =@x1j > 0 we conclude that dx11 =dc ˆ 0 and that dx12 =dc < 0. CP±SP strategy. From (7) we obtain the ®rst order condition @PCPÿSP =@x1 ˆ ‰P1 ÿ …b11 x1 ÿ K1 …Attr1 Q1 ‡ Attr2 Q2 ††Š  …@Attr1 =@x1 †Q1 ‡ ‰rP2 ÿ …b11 x1 ÿ K1 …Attr1 Q1 ‡ Attr2 Q2 †† ÿ cŠ  …@Attr2 =@x1 †Q2 ‡ ‰Attr1 Q1 ‡ Attr2 Q2 Š

Proof of Proposition 1

 ‰ÿb11 ‡ K1 …@‰Attr1 Q1 ‡ Attr2 Q2 Š=@x1 †Š ˆ 0:

If we impose the constraint x1 ˆ x2 ˆ x on the pro®t function of the DP±CP strategy given in equation (5) we obtain the pro®t function of the DP±SP strategy given in equation (8). Therefore, max PDPÿCP 5 max PDPÿCP jx1 ˆ x2 ˆ x ˆ max PDPÿSP

…13† Let h ˆ @PCPÿSP =@x1 . Then we obtain @h=@c ˆ ÿ…@Attr2 =@x1 †Q2 : We can easily show that dx1 @h=@c ˆÿ : dc @h=@x1

Proof of Proposition 2 DP±CP strategy. The pro®t function of the DP±CP strategy in equation (5) does not include the variable c (tariffs/ import cost) since production takes place locally in each country. As a result, the optimal product design resulting from the pro®t maximisation program of the DP±CP strategy given in equation (5), is not affected by c. CP±CP strategy. From (6) we obtain the ®rst order conditions @PCPÿCP =@x11 ˆ ‰P1 ÿ …b11 x11 ÿ 2K1 Attr1 Q1 †Š  …@Attr1 =@x11 †Q1 ÿ b11 Attr1 Q1 ˆ 0; …9†

 …@Attr2 =@x12 †Q2 ÿ b11 Attr 2 Q2 ˆ 0: …10† Let q11 ˆ @PCPÿCP =@x11 and q12 ˆ @PCPÿCP =@x12 . Then, we obtain @q12 =@c ˆ ÿ…@Attr2 =@x12 †Q2 :

…11†

Since q1j …x1j …c†; c† ˆ 0 8c, the chain rule gives @q1j dx1j @q1j dc ‡ ˆ 0: @x1j dc @c dc

Using (14), (15) and the assumptions that @2 P=@x21 < 0 and that @Attrj =@x1 > 0; j ˆ 1; 2, we conclude that dx1 =dc < 0. Proof of Proposition 3 DP±CP strategy. From (4) we obtain the ®rst order conditions @Pj =@x1j ˆ ‰Pj ÿ …b1j x1j ÿ 2Kj Attrj Qj †Š…@Attrj =@x1j †Qj ÿ b1j Attrj Qj ˆ 0;

j ˆ 1; 2:

…16†

From (16), it is evident that r does not affect the optimal product attribute level in either country. This is due to the decentralised nature of the DP±CP strategy. Nevertheless, the exchange rate affects the total pro®ts from country 2. @q11 =@r ˆ 0;

@q12 =@r ˆ P2 …@Attr2 =@x12 †Q2 :

…17†

We can easily show that dx1j @q1j =@r ˆÿ ; @q1j =@x1j dr

j ˆ 1; 2:

…18†

Using (17), (18), and the assumptions that @2 P=@x21j < 0 and that @Attrj =@x1j > 0 we conclude that dx11 =dr ˆ 0 and that dx12 =dr > 0. CP±SP strategy. From the proof of Proposition 2 we obtain @h=@r ˆ P2 …@Attr2 =@x1 †Q2 :

The above equation yields dx1j @q1j =@c ˆÿ ; @q1j =@x1j dc

…15†

CP±CP strategy. From the proof of Proposition 2 we obtain

@PCPÿCP =@x12 ˆ ‰rP2 ÿ …b11 x12 ÿ 2K1 Attr2 Q2 † ÿ cŠ

@q11 =@c ˆ 0;

…14†

…19†

We can easily show that j ˆ 1; 2:

…12†

dx1 @h=@r ˆÿ : dr @h=@x1

…20†

GC Hadjinicola and KR KumarÐFactors affecting international product design

Using (19), (20) and the assuptions that @2 P=@x2i < 0 and that @Attrj =@x1 > 0 we conclude that dx1 =dr > 0. Proof of Proposition 4 DP±CP strategy. From the proof of Proposition 3, let z1j ˆ @Pj =@x1j . Then, using (16) we obtain @z1j =@Kj ˆ 2Attr j …@Attrj =x1j †Q2j ;

j ˆ 1; 2:

…21†

We can easily show that dx1j @z1j =@Kj ˆÿ ; dKj @z1j =@x1j

j ˆ 1; 2:

…22†

Using (21), (22) and the assumptions that @2 P=@x21 < 0 and that @Attrj =@x1j > 0 we conclude that dx1j =dKj > 0. CP±CP strategy. From the proof of Proposition 2 we obtain @q1j =@K1 ˆ 2Attrj …@Attrj =x1j †Q2j ;

j ˆ 1; 2:

…23†

We can easily show that dx1j @q1j =@K1 ˆÿ ; dK1 @q1j =@x1j

j ˆ 1; 2:

…24†

Using (22), (23), and the assumptions that @2 P=@x21j < 0 and that @Attrj =@x1j > 0 we conclude that dx1j =dK1 > 0. CP±SP strategy. From the proof of Proposition 2 we obtain @h=@K1 ˆ 2‰Attr 1 Q1 ‡ Attr2 Q2 Š  ‰…@Attr1 =@x1 †Q1 ‡ …@Attr2 =@x1 †Q2 Š:

…25†

We can easily show that dx1 @h=@K1 ˆÿ : dK1 @h=@x1

…26†

Using (25), (26) and the assumptions that @2 P=@x21 < 0 and that @Attrj =@x1 > 0 we conclude that dx1 =dK1 > 0. References 1 Levitt T (1983). The globalization of markets. Harvard Bus Rev, May±June, 92±102. 2 Boddewyn JJ, Soehl R and Picard J (1986). Standardization in international marketing: Is Ted Levitt in fact right? Bus Horizons, November±December, 69±75.

1143

3 Douglas SP and Wind Y (1987). The myth of globalization. Columbia J of World Bus, Winter, 19±29. 4 Elinder E (1961). How international can advertising be?. The Int Advertiser, December, 12±16. 5 Kashani K (1989). Beware of the pitfalls of global marketing. Harvard Bus Rev, September±October, 91±98. 6 Porter ME (1990). The Competitive Advantage of Nations. The Free Press: New York. 7 Hayes RH and Wheelwright SC (1984). Restoring Our Competitive Edge. John Wiley and Sons: New York. 8 Oliff MD, Arpan JS and Dubois FL (1989). Global manufacturing rationalization: The design and management of international factory networks. In: Ferdows K (ed), Managing International Manufacturing. North Holland: New York, pp 41±65. 9 Schocker AD and Shrinivasan V (1974). A consumer-based methodology for the identi®cation of new product ideas. Mgmt Sci 20: 921±937. 10 McFadden D (1986). The choice theory approach to market research. Marketing Sci 5: 275±297. 11 Gruca TS, Kumar KR and Sudharshan D (1992).An equilibrium analysis of defensive responses to entry using a coupled response function model. Marketing Sci 11: 348±358. 12 Albers S and Brockhoff K (1977). A procedure for new product positioning in an attribute space. Eur J Opl Res 1: 230±238. 13 Sudharshan D, May, JH and Gruca TS (1988). DIFFSTRAT: An analytical procedure for generating optimal new product concepts for a differentiated-type strategy. Eur J Opl Res 36: 50±65. 14 Eliashberg J and Manrai AK (1992). Optimal positioning of new-product concepts: Some analytical implications and empirical results. Eur J Opl Res 63: 376±397. 15 Choi CS, DeSarbo WS and Harker PT (1990). Product positioning under price competition. Mgmt Sci 36: 175±199. 16 Bachem A and Simon H (1981). A product positioning model with costs and prices. Eur J Opl Res 7: 362±370. 17 Hout T, Porter ME and Rudden E (1982). How global companies win out. Harvard Bus Rev, September±October, 98±108. 18 Yip GS (1989). Global strategy. . .In a world of nations?. Sloan Mgmt Rev, Fall, 29±41. 19 Hadjinicola GC and Kumar KR (1994). Facility rationalization and product desing in international enterprises, Report 94-09, Department of Public and Business Administration, University of Cyprus. 20 Schwartz HR (1989). Numerical Analysis: A Comprehensive Introduction. John Wiley & Sons: New York. 21 Walters PG and Toyne B (1989). Product modi®cation and standardisation in international markets: Strategic options and facilitating policies. Columbia J World Bus, Winter, 37±44.

Received September 1996; accepted June 1997 after two revisions