|
Article Excerpt
In this paper we characterize the impact of production technology on the optimal product line design. We analyze a problem in which a manufacturer segments the market on quality attributes and offers products that are partial substitutes. Because consumers self-select from the product line, product cannibalization is an issue. In addition, the manufacturer sets a production schedule in order to balance production setups with accumulation of inventories in the presence of economies of scale. We show that simultaneous optimization of the product line design and production schedule leads to insights that differ significantly from the common intuition and assertions in the literature, which omits either the demand side or the supply side of the equation. In particular, we demonstrate that more expensive production technology always leads to lower product prices and may at the same time lead to higher quality products. Further, a less efficient production technology does not necessarily increase total production costs or reduce consumer welfare. We also demonstrate that in the presence of production technology, the demand cannibalization problem may distort product quality upward or the number of products upward, which is contrary to the standard result.
Key words: product line; segmentation; cannibalization; EOQ; scale economies; marketing-manufacturing interface
1. Introduction
Consumers seek product variety for a number of reasons (Hoch et al. 1999). First, greater variety ensures that a consumer finds exactly what he/she wants. Second, when preferences are uncertain, variety may offer an option value. Third, consumers may have an inherent desire to try different alternatives. Although there are exceptions in which variety can have a negative effect (see Gourville and Soman 2005), it is a generally accepted proposition that offering larger variety allows companies to increase both demand and market share (Kotler 2002). However, it has also been long recognized that a large number of products in the assortment is associated with diseconomies of scale and increases in production and distribution costs (Lancaster 1990, Quelch and Kenny 1994). For example, in a survey of product variety models, Lancaster (1990, p. 189) points out that "scale economies mean that the potential welfare or revenue gain from greater product variety must be balanced against the lower unit production cost with fewer variants." Hence, with the exception of a few firms (e.g., Dell) that are able to manufacture products to customer order and avoid large inventories, most companies face a tradeoff between product variety and production costs, so that product line decisions should be made with production costs in mind (Eliashberg and Steinberg 1993). (1) An intuitive corollary is that higher production costs should lead to a reduction in the number of products offered, higher prices (Bayus and Putsis 1999, Kekre and Srinivasan 1990), and lower quality.
Surprisingly, empirical literature in marketing has had mixed success in finding evidence of these effects (Kekre and Srinivasan 1990), and there are very few models that provide guidance with respect to the precise impact of production costs on product line decisions. In particular, it is often assumed in the extant literature that adding a product to the assortment simply results in a fixed cost (see Dobson and Kalish 1988, Yano and Dobson 1998). A simple example showing this approach to be problematic is that an increase in fixed cost does not affect product quality or prices (unless the increase is so high that a product is dropped), which is contrary to the common understanding that production costs should affect product quality and prices. To improve on this dimension, we propose a dynamic model in which a product line and a production schedule are set simultaneously. We consider a manufacturer who segments the market based on quality attributes of the product, and, all else equal, consumers prefer higher quality. Examples of such product lines are the same model of automobile that may come with various levels of gas-mileage performance (e.g., hybrid or conventional engine), multiple variations of the same computer with different sizes of hard drives, processors of varying speeds, chemicals of different purity, and paper of different density. Consumers are heterogeneous in their preferences for quality, and self-select from the product line. Hence, a lower quality product can cannibalize the higher quality product, an effect that the manufacturer must consider.
Consumers arrive dynamically and a firm sets a production schedule ensuring that an adequate amount of product is manufactured. Diseconomies of scale, inventory carrying costs, and fixed production costs are the most frequently cited reasons to reduce product variety (see Lancaster 1990, Eliashberg and Steinberg 1993). For example, automotive manufacturers struggle with increasing product variety that results in lower demand per model, and hence diseconomies of scale, higher inventories, and higher overhead expenses (Hayes and Wheelwright 1984). We incorporate all three of these elements, for we also believe they are important determinants of optimal product line decisions. In our model, production occurs in batches that incur fixed costs. Batch sizes have to balance these costs with inventory carrying costs incurred between batches. Such a setup is common for many manufacturing firms that have to manufacture products in advance of customer demand.
We focus on two important aspects of product line design (see Figure 1). The literature in this area generally assumes that firms produce to order (an approach we call classical). First, we analytically characterize the impact of production technology on product line design. This is done by comparing the production-to-order setup with the more elaborate model, which includes a production-to-stock setting, inventories, and economies of scale. The second dimension of our analysis is the impact of information in the presence of production technology. This is done by comparing two settings: the benchmark setting with full information about consumer preferences and the setting with asymmetric information. In the full-information case, the firm knows the preferences of individual consumers so that the cannibalization problem does not arise because the firm can segment consumers perfectly. In the asymmetric information case, the firm cannot observe the preferences of individual consumers, and hence lower quality products can cannibalize higher quality products. We demonstrate that these two factors--production technology and information--interact and have a major impact on the firm's product line decisions. Our main findings are:
* Production costs reduce a firm's propensity to offer multiple products in favor of offering fewer products when production costs are large enough. This effect, however, is moderated by the (dis)similarity of consumer segments. When consumer segments are relatively close to each other, the firm offers one composite product designed to serve multiple segments. When consumer segments are far apart, the firm also reduces the number of products, but this time through not serving consumers at the low end of the market.
* More expensive production technology (in the sense of higher relevant cost parameters) can lead the firm to offer a product line of higher (average) quality at a lower (average) price. This occurs when the more expensive production technology makes it attractive to replace segment-specific product offerings with a composite product; economies of scale make it attractive to increase the quality level, but the product must be priced relatively low in order to appeal to the lower-end consumer segment. More generally, more expensive production technology always leads to lower product prices. Hence, using product prices as proxies for production costs, as is sometimes done in empirical studies, can be problematic because these variables can be inversely related. This occurs because a more expensive production technology makes producing lower quality products attractive.
* A firm offering more products may have lower total production costs than a firm offering fewer products. This counterintuitive finding is consistent with (and may help explain) some empirical evidence on the relation between product variety and production costs (see Kekre and Srinivasan 1990). This result relies on the fact that firms may differ both in their production technology and in the markets they face. Another result is that a firm with less efficient production technology has lower total production costs. This occurs because less efficient production technology makes producing lower quality products attractive, and these products are cheaper to produce. Finally, from the consumer welfare point of view, it can be beneficial to have more expensive production technology. Again, this occurs when the more expensive production technology makes it attractive to replace segment-specific product offerings with a composite product.
* Production technology alters the effect and value of information about consumer preferences. Under full information, the product line is "efficient" in that it maximizes firm profit and social welfare (Moorthy 1984). Typically (see Moorthy 1984, Desai 2001, Villas-Boas 2004, etc.), the firm facing the cannibalization problem due to information asymmetry produces (weakly) fewer products that are of (weakly) lower quality than is efficient. Further, prices are distorted downward. (2) We show that when production technology is considered, these results may be reversed, i.e., the cannibalization problem may cause the firm to offer more products, or they can be of higher quality than is efficient. Further, prices may be distorted upward. These results occur because the cannibalization problem may lead the firm to offer a segment-specific product line when a composite product is efficient.
* Production technology may help mitigate efficiency losses due to information asymmetry. While with production-to-order there are always distortions from the efficient product line (either in the number of products or in their quality), with production-to-stock these distortions may be eliminated. Hence, the value to the firm of information about individual consumer preferences may be zero. This occurs when the production technology makes serving the lower end of the market unattractive under full information.
[FIGURE 1 OMITTED]
Overall, our results demonstrate that close attention should be paid to the interplay between the production technology and cannibalization problems associated with product line design. Simultaneous optimization of the product line design and production schedule leads to insights that differ significantly from the common intuition and assertions in the literature, which omits either the demand side or the supply side of the equation. Therefore, without clear understanding of the trade-offs involved, there is potential for managers to make serious judgement errors.
The paper proceeds as follows. Section 2 surveys the related literature. Section 3 models the product-line design problem under full information and under asymmetric information. Section 4 compares the results of these models and describes their implications. Section 5 discusses the impact of relaxing some of the modeling assumptions and provides concluding remarks.
2. Literature Survey
The literature on product line design has a long history, which is surveyed in Lancaster (1990), Eliashberg and Steinberg (1993), Krishnan and Ulrich (2001), and Ramdas (2003). The stream of work most relevant to our paper considers quality as a differentiating dimension along with price, and allows consumers to self-select from variants offered. We utilize the quality differentiation approach for two reasons. First, production and inventory carrying costs typically depend on product quality, but may not be affected by other differentiating dimensions (e.g., color). Second, we are interested in studying the interaction between the cannibalization effect and production costs.
The stream of research on quality differentiation was pioneered by Mussa and Rosen (1978), and Moorthy (1984) was the first to introduce this framework into the marketing literature. In Moorthy (1984) the monopolist offers a menu of products, with higher quality products sold at higher prices. Due to the cannibalization effect, products are priced so that only consumers with the highest valuation for quality receive their efficient quality level, while all...
|
