Application of Revenue Management to the Manufacturing Industry

Table of contents

List of abbreviations

List of symbols

List of figures

List of tables

1 Introduction
1.1 Problem statement
1.2 Objective

2 Basics of Revenue Management
2.1 Definition and Historical Background
2.2 Price- vs. Quantity-based Revenue Management
2.3 Instruments
2.4 Forecasting
2.5 Measuring RM performance

3 Requirements for quantity-based RM
3.1 Market
3.2 Company

4 Application to the Manufacturing Industry
4.1 Definition of manufacturing
4.2 Applicability in Service Industries vs. Manufacturing Industries
4.2.1 Market
4.2.2 Company
4.2.3 Differences between the service industry and manufacturing
4.3 In-depth look at MTO and ATO
4.3.1 Implications for the use of RM methods
4.3.2 Static models for bid-price generation
4.3.3 Self-adjusting bid-price
4.4 Current use of RM in manufacturing industries

5 Case Study
5.1 Setting of the case study
5.2 RM-fitness
5.3 Introduction of the Test Case
5.4 Implementation

6 Conclusion

7 Future Trends and Challenges

References

List of abbreviations

illustration not visible in this excerpt

List of symbols

illustration not visible in this excerpt

List of figures

Figure 1: Pricing in a non-segmented market

Figure 2: Pricing in a segmented market

Figure 3: Bid-prices, booking limits and protection levels

Figure 4: Three-dimensional view of demand

Figure 5: push-pull-boundaries

Figure 6: distribution of contribution margin in the steel industry over the course of one month

Figure 7: Revenue management system in the Ford Moter Company

Figure 8: distribution channels of steel products

Figure 9: product groups and product families of strip steel

Figure 10: implemented test case

Figure 11: order acceptance process

List of tables

Table 1: Fulfillment of requirements for quantity-based RM application in different industries

Table 2: Normalized relative capacity consumptions [h/order]

Table 3: Normalized order characteristics, .

Table 4: Resulting contribution margin as percentage of the ex-post optimal solution

1 Introduction

1.1 Problem statement

Revenue management (RM) is the umbrella term for a set of strategies, tactics and instruments aiming at the maximization of yield by allocating a company's capacity to different customers at different prices. Due to its great success, the application of revenue management is widespread nowadays. But as the origin of RM lies in the airline industry, this is still the sector of its main application. Service industries such as hotels, car-rentals or internet service providers which share the same characteris­tics as the airline industry (e.g. fixed capacity and a highly uncertain demand) dis­covered quite early the potential of RM. Consequently, they were the first to adopt RM strategies.^[1] Retailers, broadcasting industries and companies of the energy sec­tor have followed lately.

The core concept of RM becomes clear, considering the economics of RM (Cross 1997, p.73ff): The downward-sloping demand curve (figure 1) shows the number of units of a certain product which are sold at a certain price.

illustration not visible in this excerpt

Figure 1: Pricing in a non-segmented market

Trying to optimize the created revenue, the vendor will try to pick the combination of price and sold units so that their product is as big as possible. (The size of the pro­duct of price and sold units can also be illustrated as the area of the right quadrilater- al under the demand curve.) But by doing so the vendor could obviously have sold some of the units at a higher price than what he had chosen, since some of the cus­tomers were willing to pay more than what he had charged. In addition, he could have generated more revenue if he had offered the unit at a lower price, because some customers were not willing to pay the price he had picked. Evidently it would be optimal for the vendor to charge every customer the price he is actually willing to pay, since this would maximize the vendor's revenue. However, since it is unrealistic to expect to capture this perfect one-to-one pricing, the vendor should be able to divide his market in different segments, estimate the average price they are willing to pay and still make a profit using that average price (Phillips 2005, p.75ff). This effect is shown in figure 2.

illustration not visible in this excerpt

Figure 2: Pricing in a segmented market

As RM originates from the Airline Industries, the term revenue management was coined by the existing cost structures in those fields. The cost structures are charac­terized by relatively high fixed costs (for planes, hubs etc.) compared to low marginal costs (extra fuel, meals etc.) for an additional sold unit, making the sale of an addi­tional unit profitable for a wide range of prices (Rehkopf 2006, p.43). In this scenario, the maximization of revenue is almost congruent with the maximization of yield which led to the term revenue management.

Taking into account that some of the common characteristics can also be found in manufacturing industries, RM is considered to be suitable for them as well. However, research and implementation in this sector are still in the very beginning stages. One reason for this is the reduced effect on profit compared to service industries, resulting from higher variable costs. Furthermore, the necessity to change assumptions and

loosen restrictions when adapting RM to manufacturing industries leads to difficulties when applying these concepts (Chiang et al. 2007, p.116).

This paper shows possibilities and restrictions for the application of RM in the manu­facturing sector.

1.2 Objective

In Chapter two a definition of RM and a brief introduction to its historical background will be given. Moreover, the difference between price- and quantity-based RM will be explained and an introduction to the most commonly used instruments in revenue managements systems (RMS) will be provided. In addition, an insight into forecasting and its inherent problems and difficulties will be given. Moving on to chapter three, requirements concerning the market and the company that need to be fulfilled in or­der to perform a RMS successfully are identified and explained. Those requirements are only referring to quantity-based RM. The next chapter provides an analysis of whether the concept of RM can be used successfully in the manufacturing industries. This is accomplished by first defining manufacturing and then examining whether the requirements are fulfilled and instruments are applicable. Furthermore, make-to-order (MTO) and assemble-to-order (ATO) concepts are given an in-depth look before im­plications for the use of RM methods are presented. Finally, network models for bid-prices and self-adjusting bid-prices are explained. Chapter four concludes with an overview of the current use of RM in the manufacturing industries. In order to show the application of some of the instruments and their implementation, a case study is presented at the end of the paper. At first, the given situation and industry is intro­duced and the test case explained before the implementation of the model is shown. The conclusion of the paper provides critical reasoning of the topic and an outlook for further research concerning the effectiveness and usefulness of RM.

2 Basics of Revenue Management

2.1 Definition and Historical Background

The origin of revenue management can be found in the airline industry. When the Airline Deregulation Act loosened price-restrictions in the USA in 1978, the Airlines (Airline Companies) were henceforth allowed to freely set prices for flights. Therefore, new companies could enter the market. These, mainly low-cost, airlines enticed many customers, especially non-business travelers, away from the old established airlines. To counteract this process, prevailing airlines such as American Airline ea­gerly started to develop strategies to recapture lost passengers. With marginal costs near zero, they could give leftover seats away for less. Preconditions are that they were able to estimate the amount of otherwise unused seats (surplus capacity) and effectively prevent solvent customers like business travelers from taking the low-cost fares. At that time, strategies covering those issues were referred to as "yield mana­gement" which is now mostly known as "revenue management". Vast progress was made in the field of RM and more accurate models for specific problems and other branches were evolved.^[2]

The wide range of applications and research topics for RM lead to a large number of different definitions. These depend largely on the application and focus on either the used instruments or characteristics of the application. An overview of definitions is given in Kimms and Klein (2005, p.4f.) and Rehkopf (2006, p.40f.) among others. In the following discussions, a general definition suitable for RM on the manufacturing sector is used. It is given by Pak and Piersma (2002, p.1): "Revenue management can be defined as the art of maximizing profit generated from a limited capacity of a product over a finite horizon by selling each product to the right customer at the right time for the right price."

2.2 Price- vs. Quantity-based Revenue Management

Generally two different types of RM can be distinguished. Quantity-based and price-based RM. Regarding existing literature on RM, it turned out that in many papers, the problem of distinguishing price- and quantity-based RM has been avoided. Often not even a classification of the particular work is given. To prevent incorrect interpretation here, an overview of the differences as stated by Talluri and van Ryzin (2005, pp. 176-178) follows.

The original application of RM is of quantity-based character. Here decisions are made as to what quantity of an already priced product is to be offered (Rehkopf 2006, p.37). An example for this is the traditional airline industry, where fares for va­rious categories are set way in advance of the actual flight to be printed in catalogues etc. As a result, a variation of the price to influence demand is not possible after­wards. In order to maximize revenue, the allocation of available seats to the different travel-categories and such to various price-classes is an adequate tool (Talluri and van Ryzin 2005, p.176). Quantity-based RM can be used in many industries with flex­ible supply and price-commitments.

In prices-based RM, the price of a product has to be variable and is used to manage demand. Typical industries using price-based RM are retailers. Often the variation of the price of a product is cheaper than a variation of the capacity.

To summarize, the differences result from whether price or capacity can be varied, thus which is the decision variable. A strict segmentation in which industries price-based RM and in which quantity-based RM should be used is not possible. In some aspects price-based revenue management might be more profitable than quantity-based revenue management since it is increasing the profit by increasing the price rather than by limiting capacity (Talluri and van Ryzin 2005, p.177). But as Gallego and van Ryzin (1994) mention, decisions should not be made based on ideas of dy­namic pricing alone. Mostly, a mixture of both pricing and allocation schemes is prac­tical to receive the best revenues and optimal results.

2.3 Instruments

As discussed above, a differentiation can be made between price- and quantity-based RM, each containing different instruments to find the optimal set of controls (Talluri and van Ryzin 2005, p.18).

Starting with quantity-based RM some of the most common and important instru­ments are

- single resource capacity control,
- network capacity control and
- overbooking

which are discussed below.

The concept of single resource capacity control contains "allocating capacity of a resource to different classes of demand" (Talluri and van Ryzin, 2005, p.27). Here it is important to point out that only one resource, e. g. a single leg flight or only one machine, is taken into account (in comparison to network capacity control which will follow later).

For allocating the capacity of a resource to various classes of demand, different me­thods and models are needed. Since there is not enough space to explain all of those, we will only give the example of booking limits. Bid-prices will be explained in the context of network capacity control since the single-resource version is similar to the network one. Hence repetitions of contents will be avoided.

Booking limits are control mechanisms in single resource capacity control which de­fine how much of the total capacity can be used by one segment or class (Rehkopf 2006, p.53). If the assigned capacity is sold, all demands are rejected, even if there are more resources available at a lower price-segment class (Kimms and Müller-Bungart 2003, p.4). Due to this inefficiency, the specification of nested booking limits is used as well in order to give another opportunity of allocating capacity. In this concept, resources "that are available for sale to a particular booking class are also available to bookings in any higher fare booking class, but not the reverse" (McGill and van Ryzin 1999, p.249). Consequently, demands from higher price-segments classes have access to the contingent of the lower price-segment classes.

In reality, usually more than one resource has to be taken into account which leads to network capacity control. Whenever customers buy resources in bundles, network capacity control has to be used instead of single resource capacity control. This is important for the manufacturing industry since almost all products require the use of several machines. Because of interdependencies and that it is necessary to have capacity available at all resources, optimizing each resource separately does not produce the optimal result (Rehkopf 2006, p.71). Therefore, jointly managing all ca­pacity controls on all resources becomes inevitable (Talluri and van Ryzin 2005). In addition, it has been shown that heuristic methods which try to assign prices to each resource are less effective than methods which examine the network as a whole. The reason for this is the information that gets lost (Rehkopf 2006, p. 71). Network me­thods also procure considerably higher revenue benefits. However, inherent with that positive result are significant implementation and methodological challenges (Talluri and van Ryzin 2005, p.82f.) as well as a difficult and expensive transition from single-resource to network RM.

Just as single resource capacity control has a specific set of control mechanisms so too does network capacity control. Talluri and van Ryzin (2005, p.83) give some im­portant criteria for choosing the right control mechanisms:

- technological constraints imposed by the distribution system,
- revenue performance achievable by the method and
- overall robustness of the control scheme.

The major categories of network controls are partitioned booking limits, virtual nest­ing controls and bid-price controls. Most of them are basically the same as for single-capacity control but were expanded and adjusted to be able to apply them to net­works.

Partitioned booking limits allocate a certain amount of capacity to each individual product on each resource. Those portions do not overlap and may not be used by other products. Due to the great variety of products, the number of fixed allocations to resources becomes immensely large and the allocations to the segments are only fragments. Because of this situation using those partitioned booking limits can result in huge inefficiencies. Therefore, partitioned booking limits are applied very rarely in practice (Talluri and van Ryzin 2005, p.84).

[...]

^[1] Refer to Talluri and van Ryzin (2005) and Chiang et al. (2007) for an outline of RM applications.

^[2] Chiang et al. (2007) provide a comprehensive overview of research on RM.