Revenue management optimization for make-to-order and engineering-to-order manufacturing context

Revenue management optimization for make-to-order and engineering-to-order manufacturing context Jose V. Cavero a , M.M.E. Alemany b,* , Ana Esteso b , Alfredo Gim´ enez c a Industrial Engineering School, Universitat Polit` ecnica de Val` encia, Camino de Vera s/n, 46022, Valencia, Spain b Research Centre on Production Management and Engineering (CIGIP), Universitat Polit` ecnica de Val` encia, Camino de Vera s/n, 46022, Valencia, Spain c Factor Ingeniería y Decoletaje, S.L., C/ Regadors 2 P.L, Campo Anibal, Puçol, 46530, Valencia, Spain ARTICLE INFO Keywords: Revenue Management (RM) Optimization Mixed Integer Linear Programming (MILP) Order promising MTO ETO ABSTRACT The order promising process for Make-To-Order (MTO) and Engineering-To-Order (ETO) companies usually implies not only verifying delivery quantity and due date feasibility but also negotiating prices with clients derived from product customization. The existence of variable operation costs due to process selection and volatile raw material prices, emphasizes the need for precise supplier choice and cost estimation for pricing. These decisions fall within the scope of Revenue Management (RM), which is addressed in this paper through a novel optimization model for companies with hybrid MTO/ETO manufacturing strategy and even extendable to Make-To-Stock (MTS) one. A Conceptual Framework is proposed to characterize the RM problem, systematically review existing literature, and justify the novelties of this research. The model is tested on a metal-mechanic company with characteristics not previously modelled, and evaluated under different scenarios, demonstrating its effectiveness providing a price acceptable for the customer as well as profitable for the company. 1. Introduction Order management has become a key process for companies, as it directly affects customer satisfaction and the likelihood that customers will place new orders (Ould, 1995). In this context, the decision on whether to accept or reject customer orders during the order promising process is crucial to ensure the profitability and continuity of companies (Spengler et al., 2007). Order promising refers to the set of business activities triggered to provide a response to customer order requests (Grillo-Espinoza et al., 2019). The concept of Revenue Management (RM) arises when the order promising process also involves setting prices. RM encompasses the strategies, tactics, and tools that enable revenue maximization by allocating a company’s capacity to different customers (Zatta, 2016, Talluri and van Ryzin, 2004). Originally intended for the service sector, particularly the airline industry, this concept has been extended to various stages of manufacturing supply chains, including distribution (Raza et al., 2018) and sales (Chen and Hsu, 2017). Today, there are also successful RM applications in the manufacturing stage within Make-To-Order (MTO) and/or Engineering-To-Order (ETO) environments (Harris and Pinder, 1995, Rehkopf and Spengler, 2005). This is because they share common characteristics with RM, such as capacity control due to the inflexible and perishable capacities, dynamic pricing, the possibility to produce with the order already confirmed, or product customization (Müller-Bungart, 2007). However, significant differences exist in the application of RM to MTO/ETO environments (Lohnert and Fischer, 2019). In MTO/ETO companies, unlike the service sector, it cannot always be assumed that an increase in revenues directly leads to an increase in profits due to the significant variable costs associated with the products. Therefore, it is necessary to replace the concept of revenue with that of contribution margins (Spengler et al., 2007). Moreover, in order to apply RM, it should be possible to reject customer order proposals and/or adjust prices, allowing for the customization of not only the products but also, to some extent, the prices (Hintsches et al., 2010). In addition, for RM to be most effective, production capacity in MTO environments should be considered as perishable inventory (Kimes, 1989). Finally, while late and early deliveries are penalized, they are still possible. The most typical RM challenges in MTO/ETO include rigid shortterm production capacities, a broad product catalogue, and pronounced fluctuations in demand and profitability (Hintsches et al., 2010). These fluctuations are usually driven by volatile and uncertain * Corresponding author. E-mail addresses: [email protected] (J.V. Cavero), [email protected] (M.M.E. Alemany), [email protected] (A. Esteso), [email protected] (A. Gim´ enez). Contents lists available at ScienceDirect International Journal of Production Economics journal homepage: www.elsevier.com/locate/ijpe https://doi.org/10.1016/j.ijpe.2025.109715 Received 31 March 2024; Received in revised form 6 December 2024; Accepted 19 June 2025 Int. J. Production Economics 288 (2025) 109715 Available online 24 June 2025 0925-5273/© 2025 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC license ( http://creativecommons.org/licenses/bync/4.0/ ). variable costs, often resulting from highly unstable raw material purchase prices and availability. In this context, achieving the best offers from potential suppliers in terms of quantity, price and delivery-date becomes crucial for remaining competitive in the market. This is even more relevant if alternative raw materials are available for manufacturing the same final product. Moreover, ETO manufacturing strategies differ from MTO, as they may require either new or adapted designs to a greater or lesser extent, as well as the modification of the manufacturing process, including the selection of raw materials, tools, and machines used in production. Indeed, the distinction between MTO and ETO is based on the location of the customer order decoupling point (CODP) along the production flow: upstream the CODP the production activities are driven by forecasts and inventory exists, while downstream the CODP no inventory exists since manufacturing activities are order-driven (Wikner and Rudberg, 2005). The CODP location gives rise to different manufacturing strategies (Cannas et al., 2020) ranging from pure standardisation, such as make-to-stock (MTS), to pure customization, such as ETO. These strategies also include segmented standardisation, such as delivery-to-order (DTO), customised standardisation, such as assembly-to-order (ATO), and tailored customization, such as MTO. However, to more precisely capture the interactions between production and engineering (Wikner and Rudberg, 2005), these initial production-based CODPs were complemented with engineering-based CODPs in a two-dimensional framework (2D-CODP). The goal is to distinguish among engineering activities driven by orders and those driven by forecasts. This 2D-CODP framework assumes that engineering and production are different flows of activities that can be decoupled independently, making necessary to carefully considering the interfaces between them (Cannas et al., 2019). Therefore, ETO companies can adopt different order-fulfilment strategies by deciding how to balance order-driven and forecast-driven engineering and production activities according to their priorities and the trade-off among the delivery time, price, flexibility, uniqueness and reliability. Cannas et al. (2020) affirm that a gap still exists between theory and practice when analysing the coordination between engineering and production to support ETO companies in adapting their order-fulfilment strategies to the market. They state that recent studies have primarily focused on customization choices in the engineering process, with little attention given to its interactions with the production process. Similarly, other authors (Fortes et al. 2023a, 2023b) identify production scheduling, production planning and control and capacity planning as some of the most critical issues for ETO companies from both academic and industrial perspectives. Additional challenges include batch size, pricing, procurement and closer collaboration between engineering and production, all them addressed in this research. Bejlegaard et al. (2021) suggest that potential strategies to tackle these challenges include lean and agile management practices, enhanced communication and coordination among different departments, and the implementation of advanced tools and technologies to boost process efficiency. Cannas and Gosling (2021) analyse a wide range of ETO challenges, including the need for future research on strategies to improve supply chain management performance in ETO situations. They also identify the management and implementation of transitions between decoupling configurations as an existing challenge. This is relevant since it is usual the co-existence of different CODP locations in the same company, which may also change over time. For instance, a novel product can be managed using an ETO strategy the first time it is requested by a customer. However, for additional orders of the same product, the MTO strategy can be adopted since no engineering activities are necessary due the availability of its previous design. Furthermore, if this product becomes regularly demanded, the company might decide to produce it to stock (MTS). Despite all of this, the inherent complexity of hybrid MTO/ETO environments has received very little attention in the literature (Barbosa and Azevedo, 2018). Afnes et al. (2021) state that the CODP positioning not only defines the type of ETO system but also the degree of uncertainty to be encountered and managed. They distinguish between two ETO types: innovate-to-order (ITO) and redesign-to-order (RTO). ITO systems generate innovative projects meanwhile RTO systems develop new products from existing designs of previous projects that are modified. For this reason, ITO systems require high engineering workload meanwhile RTO systems are executed with a minimum number of engineering hours in order to expand the MTO solutions with slightly customer designs for a sensitive price market. Both them are subject to high levels of uncertainty although different in nature (Afnes et al., 2023): while ITO systems mainly have uncertainties in engineering, RTO systems mainly present uncertainties in production. They also should face uncertainties in demand, internal processes, supply and their control systems by defining tactics and support tools for their mitigation. In view of the above, the aim of this paper is to support manufacturing companies with a hybrid ETO/MTO strategy during the revenue management process by means formulating a new revenue management model and conducting experiments for its validity to reduce the gap between theory and practice and to address some of the main challenges and gaps detected in the literature. Therefore, the contribution of this research is manifold. From the qualitative perspective, a Conceptual Framework (CF) for characterizing the RM in hybrid MTO/ETO companies is proposed, taking into account novel characteristics such as: a) alternative raw materials, tools, and machines for manufacturing each product; b) various raw material and tool offers from several suppliers that differ in costs and delivery dates; c) limited useful life of tools, d) raw materials assignment to warehouse locations with incompatibilities; and e) simultaneous checks of different levels of uncommitted availability (MATP, ATP, CTP and DTP). From the quantitative viewpoint, a novel mixed-integer linear programming (MILP) model is developed to optimize the joint decision-making process regarding order acceptance/rejection, pricing, due-date setting, production scheduling (lot sizing and loading), allocation of warehouse positions to raw materials, and selection of offers from different suppliers balancing the different costs incurred. Finally, managerial insights are provided to improve the coordination among engineering, production, procurement, and sales activities for better performance of hybrid MTO/ETO environments and to mitigate different sources of uncertainty by showing the utility of the RM optimization model for companies. To the authors’ knowledge, no previous studies have included all the above features. The model has been validated through a real application in the metal-mechanic sector, although its parameterized formulation enables its application across other sectors. In view of this, this article makes research contributions in the field of RM optimization applied to the order promising process for MTO/ETO companies, reducing the gap between theory and practice. The methodology adopted in this research combines both, theoretical and empirical insights in a back-and-forth direction (Fig. 1): 1) First, the conceptualization of the context and the aim of this research are described based on the literature study of theorical aspects related to the OPP and RM, conceptual frameworks and empirical observation; 2) second, a preliminary RM conceptual framework (RM-CF) along with its main dimensions are defined; 3) then, based on this first RM-CF, a systematic literature review of RM optimization mathematical models was performed which, in turn, allowed for: a) the refinement and enhancement of the preliminary RM-CF to the final RM-CF version in an iterative way, b) the identification of the extent to which the main dimensions of the problematic under study have been previously addressed by existing literature on the topic, c) the detection of existing gaps, d) the characterization of the problem addressed in this research, and e) its contributions to the literature by comparing it with detected gaps; 4) After that, the detailed problem description based on the previous characterization of the problem is performed; 5) afterwards, the MILP model formulation to optimally solve the characterized problem is developed. Once formulated, 6) computational experiments are designed including: J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 2 a) the selection of a metal-mechanic company as a representative case study of the problem addressed since it presents its key features, b) the collection of required data for executing the model and c) the definition of experimental scenarios in order to analyse the model behaviour in different situations. The experimental design has enabled 7) the model implementation and validation, also useful to refine the formulation of the initial model and to deepen its use in industrial cases. Finally, 8) the analysis and discussion of the results obtained from both theoretical and practical perspective are presented and 9) managerial implications and recommendations are determined. The rest of the paper is structured according to the research methodology followed. Section 2proposes the Conceptual Framework to characterize RM in MTO/ETO environments (RM-CF). This CF is used to analyse, in a structured manner, the most relevant RM mathematical models related to this work and to identify the main existing gaps, highlighting those covered by this paper. The problem under study is described in Section 3, while the proposed MILP model (RM-model) for solving it is detailed in Section 4. The case study and the computational experiments to validate the model and its results are analysed in Section 5. In Section 6, guidelines are provided in order to implement successfully the proposed model in an industrial setting. Section 7provides a discussion about managerial insights. Finally, Section 7draws conclusions and offers future research lines. 2. Structured literature review This section reviews the most relevant mathematical models to support the RM optimization in companies with MTO/ETO manufacturing strategies, based on previous methodologies (Seuring and Müller, 2008; Grillo et al. 2016; Lorente-Leyva et al., 2024). First, a preliminary Conceptual Framework (CF) to characterize the problem under study is proposed. Then, a structured literature review is conducted based on this framework, allowing for the iterative refinement of the preliminary CF until its final definition. The results of the review help to identify existing gaps, being those covered in this paper our contribution. 2.1. Conceptual framework for RM based order promising mathematical programming models The proposed CF is inspired by Grillo-Espinoza et al. (2016) but has been adapted to suit RM in MTO/ETO companies, making it different and therefore a contribution of this paper. The CF is composed of four categories (Fig. 2): network characteristics, order characteristics, process characteristics, and model characteristics. Each category includes different dimensions, with different values to be chosen for the specific study case characterization. The Network characteristics category contextualizes the supply chain environment through the sector, the SC stages covered and the production strategy used to meet the needs of customers. The Order characteristics category covers the customizable features in order proposals: number of items, their requested quantity and delivery date, as well as other customizable aspects such as partial deliveries, early and late deliveries, or substitute materials. The Process characteristics category addresses the aspects related to the execution mode RM, number and type of orders in each execution and the customer segmentation/prioritisation adopted. Verification during the RM process that availability levels are not enough to those in the order proposals is critical to guarantee the reliability of the delivery date committed to the client (Alemany et al., 2008). Different uncommitted availability levels exist: ATP (Available-To-Promise) related to final products, MATP (Material Fig. 1. General overview of the research methodology adopted. J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 3 Available-To-Promise) to raw materials, CTP (Capable-To-Promise) to production capacity and, finally, DTP (Deliver-To-Promise) to distribution resources (Alarc´ on et al., 2007). Additionally, it may be necessary to trace the remaining lifetime of tools in order to be compared with that required to process the customer orders. Finally, the Model characteristics category identifies the modelling techniques, the objectives pursued, the nature of the optimization function, the resolution method adopted and its validation. Last but not least, the output (network decisions) of the mathematical programming model once solved should be analysed. 2.2. Structured literature analysis and contributions of this study The scope of the literature review was limited to studies developing mathematical models for the RM optimization in ETO/MTO. Relevant research was identified by searching in the Scopus and WoS databases, using in the TITLE field the following terms: "bid pric*" or "pricing" or "cost management" or "profitable-to-promise" or "revenue management" or "yield management" AND "make-to-order" or “engineer*-to-order” or “steel” or “metal*”. The search provided 97 and 115 papers in Scopus and WoS, respectively, with 143 of potentially relevant papers, Fig. 2. CF for RM based order promising mathematical programming models. Table 1 Network characteristics. Ref Sector SC stages Productive strategy Metalmechanic Other sectors Unspecified Supplier Provisioning logistic Processor Distributor ETO MTO ATO MTS Easton and Moodie (1999)   X  X  X  Dobson and Yano (2002)   X  X  X  Defregger and Kuhn (2007) X  X X   XX Spengler et al. (2007) X   X X   X  Ebadian et al. (2008)   X X X  X  Liu and Liu (2009)   X  X  X  Hintsches et al. (2010) X    X  X  Martínez &Arredondo (2010)   X  X  X  Chaharsooghi et al. (2011) X   X  X  Kalantari and Ebadian (2011)   X X X  XX Li et al. (2012)   X  X  X  Volling et al. (2012) X    X  X  Xu and Zhang (2017) X    X  X  Garmdare et al. (2018) X   X  X  Lohnert &Fischer (2019)   XX X   X  Abedi and Zhu (2020)   X X X  XX Percentage (%) 25 18,75 56,25 6,25 18,75 100 0 0 100 0 18,75 This paper X X X X X X X   J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 4 excluding duplicates. After reviewing them, the articles that did not propose any optimization mathematical model for the RM in ETO and/ or MTO were discarded and three cross-ref were included, leaving 16 articles. These papers jointly with our proposal were analysed in a structured way based on the CF dimensions (Tables 1–7) in order to detect gaps and highlight our research contribution. As can be observed from the analysis (Table 1), the majority of the papers (56.25 %) do not specify the sector, the 25 % are developed for the metal-mechanic sector, and the remaining 18,75 % for other sectors. All papers include the processor stage, with three of them (18.75 %) also covering the supply logistics and three more (18.75 %) the supplier stage. While all studies focus on MTO environments, only three also address MTS strategies simultaneously. No paper addresses ETO, neither alone nor in conjunction with other strategies. This paper is applied to the metal-mechanical sector, although it could be used in any other manufacturing industry with similar characteristics. The optimization model of this research was firstly developed for companies manufacturing ETO and/or MTO products. However, the inherent production features addressed can result in excess final product over customer demand, building up a certain final product inventory also included in the model. This aspect makes our formulation valid for MTS as well, that can be used in case a transition to production to forecast of certain products was decided by the company. At this point, it is important to note that the RM-model of this paper does not consider the engineering workload required for developing new designs or adapting existing ones. Instead, the proposed model assists in reducing the number of engineering hours required for an order. This time saving is due to the fact that the RM-model supports the automated optimization of several decisions during the design process of new/ adapted products for ETO or existing designs for MTO, such as: order pricing, the allocation of production to machines and the selection of raw materials and tools type from different suppliers offers based on their delivery times and costs. To find the optimal value of all these interconnected decisions is a complex and very time consuming, even more without any support. This is of relevance for both, ETO and MTO, where raw materials are subject to a high market price volatility and uncertainty in their availability over time. Due to this variable and uncertain procurement and production costs, order pricing when promising orders becomes a critical aspect to ensure certain profitability for the company. As a consequence, the necessity of properly coordinate procurement and sales (pricing) activities with engineering and production becomes another priority in the problem addressed. There is a certain parity between the studies where the customer order proposal only contains one item (43.75 %) and those with several items, as in this paper (56.25 %) (Table 2). In all cases, the quantity requested for each item is specified by the customer as in this paper. In the 12.5 % of the papers, the delivery date is not specified in the order proposal, in the 25 % the delivery date is set by the customer as a range or time window while in the remaining 62.25 %, it is a single value also proposed by the customer (as in this model). Regarding the flexibility the client is willing to assume, only Defregger and Kuhn (2007) model partial deliveries (Table 3). Approximately half of the papers do not allow delayed and/or advanced deliveries, similar to this paper, which goal is to deliver the customer order just in time. No paper uses substitute materials when promising orders, except this study, constituting a distinctive factor and a contribution. All the papers are executed in an offline manner (Table 4), mainly due to the high number of checks and calculations required to estimate the delivery date, the cost and the price of orders for MTO/ETO products. Furthermore, a minority of the models (25 %) respond individually to each request, as is the case of our model. The remaining 75 % execute the RM process in batches, after considering several orders that arrive within a certain time interval. Moreover, customer segmentation is applied in five papers (31.25 %). Almost all studies, except one, consider only order proposals, not allowing any change in allocation decisions for either committed orders from the past or those in currently in the process of manufacturing (Table 5). Indeed, Garmdare et al. (2018) is the only study that considers all three types of orders, offering the possibility of re-sequencing firmed orders if the current ones improve the RM objectives. All the articles consider the production capacity available to promise which is consistent with the MTO strategy, while only five model ATP, two model MATP and none model DTP as our model. Only one paper, Defregger and Kuhn (2007), in addition to our paper, checks all three availability levels simultaneously. Moreover, our model not only considers MATP, ATP, CTP, DTP, but also the availability of tools’ lifetimes in order to ensure that production can take place on the machines. Regarding the model characteristics (Table 6), all the papers considered have a single objective function, mainly profit maximization, except for two papers and this proposal. Concerning network decisions, all papers perform the order assignment (loading), 43.75 % set the price, 25 % also perform the sequencing, 31.25 % decide the delivery date, and 25 % manage the purchase of raw materials. As novelties, the proposed model in this paper incorporates lot-sizing and loading decisions with substitute materials, the storage allocation of raw materials subject to incompatibilities, and the purchase of tools and raw materials Table 2 Order proposal characteristics: customizable aspects. Ref Items in order Specified quantity in the order Specified delivery date in order One Several Value Unspecified Value Range Unspecified Easton and Moodie (1999) XXX  Dobson and Yano (2002) X X    X Defregger and Kuhn (2007) XXX  Spengler et al. (2007) XXX  Ebadian et al. (2008) XXX  Liu and Liu (2009) X X   X Hintsches et al. (2010) XXX  Martínez and Arredondo (2010) X X X  Chaharsooghi et al. (2011) X X   X Kalantari and Ebadian (2011) X X X  Li et al. (2012) X X    X Volling et al. (2012) X X X  Xu and Zhang (2017) XXX  Garmdare et al. (2018) XX  X Lohnert and Fischer (2019) X X   X Abedi and Zhu (2020) X X X  Percentage 43,75 56,25 100 0 62,5 25 12,5 This paper X X X  J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 5 considering various offers from different suppliers. As regards the model type (Table 7), the 37.5 % of the papers develop Linear Programming (LP) models and 43.75 % Mixed Integer Linear Programming (MILP) models. Garmdare et al. (2018) presents two different models, one more basic in Non-Linear Programming (NLP) and another in Mixed Integer Non-Linear Programming (MINLP). Finally, in the “other” category, we find two papers: Defregger and Kuhn (2007) uses a Markov decision process for order acceptance, and Li et al. (2012) apply a Semi-Markov Decision Problem (SMDP) and develop a Q-learning algorithm (QLA) based on reinforcement learning. Once the model is formulated, 62.5 % of the works apply an exact resolution method, providing the optimal solution, while 56.25 % use heuristics that find satisfactory solutions. Two papers combine exact and heuristic solution methods, whereas no metaheuristics or matheuristics have been found. Model validation is mainly carried out through a numerical example (68.75 %), followed by a realistic case (25 %) where data is estimated from reality. Only one article implements a real application (Garmdare et al., 2018), as in our case. Based on the literature analysis performed, it can be stated that, to our knowledge, no existing optimization model simultaneously contemplates all the features of our proposal. Furthermore, some model characteristics are scarcely addressed or have not been previously considered at all, such as: a hybrid ETO, MTO, and MTS production strategy; substitute materials; the joint consideration of MATP, ATP, CTP, DTP, and tool lifetime; the purchase of tools from suppliers; and the storage allocation subject to incompatibility restrictions. 3. Problem description In this section, the main features of the RM problem are structured into three main blocks: the inputs, the RM process itself, and its outputs. 3.1. Main input characteristics The order proposals arriving at the company consist of one or several order lines with a common due date proposed by the client. For each order line, the customer specifies the product and its requested quantity. Accepted historical prices for the same final product and customer for MTO products, or similar designs for new ETO products, as well as market trends, are taking into account when deciding the final order price. The production strategy is assumed to be MTO and/or ETO. For MTO products, the possible raw materials, machines, and tools for processing them are known, although production will not begin until the order is committed. For ETO products, additional amount of design work is necessary whose magnitude will be dependent on the degree of design customization and, therefore, the engineering CODP location. Although the management of this engineering workload is outside the scope of this paper, the RM-model can be used for ETO products during the evaluation of several design alternatives, including the selection of potential materials and tools to use, as well as the definition of the production process. Trough the he RM-model the solutions are improved and considerable time is saved. Once the design is completed, the outputs regarding alternative raw materials, operations, and the potential machines to perform each one, are also assumed to be known for ETO products. These then become the inputs to the RM-model, being treated equivalently to MTO products during the RM process. For this reason, in Table 3 Order proposal characteristics: flexibility in delivery. Ref Partial delivery Delayed delivery Advanced delivery Substitute materials Yes No Yes No Yes No Yes No Easton and Moodie (1999) X X   XX Dobson and Yano (2002) XXXX Defregger and Kuhn (2007) X  XXX Spengler et al. (2007) XXXX Ebadian et al. (2008)   XX  X Liu and Liu (2009) X X X  X Hintsches et al. (2010) XXXX Martínez and Arredondo (2010) XXXX Chaharsooghi et al. (2011) X X   XX Kalantari and Ebadian (2011) X X X  X Li et al. (2012) XXXX Volling et al. (2012) XXXX Xu and Zhang (2017) XXXX Garmdare et al. (2018) X X X  X Lohnert and Fischer (2019) X X   XX Abedi and Zhu (2020) X    XX Percentage 12,5 81,25 43,75 50 25 68,75 0 100 This paper XXX X  Table 4 RM process characteristics: execution mode, number of orders and customer segmentation. Ref Execution mode Number of orders Customer segmentation Online Offline Single In batch Yes No Easton and Moodie (1999) XXX Dobson and Yano (2002) XXX Defregger and Kuhn (2007) X X   X Spengler et al. (2007) XXX Ebadian et al. (2008) XX X  Liu and Liu (2009) XXX Hintsches et al. (2010) XXX Martínez and Arredondo (2010) XXX Chaharsooghi et al. (2011) XX X  Kalantari and Ebadian (2011) XX X  Li et al. (2012) X X   X Volling et al. (2012) X X   X Xu and Zhang (2017) XXX Garmdare et al. (2018) XXX Lohnert and Fischer (2019) X X X Abedi and Zhu (2020) XX X  Percentage 0 100 25 75 31,25 68,75 This paper X X   X J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 6 the following, both types of products can be regarded as equal, also along the experimental design. Each product has an “ideal” raw material that constitutes the first option when manufacturing the product. However, if there is not enough availability of the “ideal” raw material, it is possible to use “substitute” materials, which may lead to increased production costs and/or processing times. This option can be also adopted to reduce stock levels and prevent the obsolescence of low-rotation raw materials, which are grouped into families for storage purposes. Moreover, several offers from different suppliers are evaluated for the acquisition of each raw material involved in a customer order proposal. An offer will include the price, delivery time, and, if applicable, the minimum order quantity or lot size. The company can select one or several offers to procure the required quantity of the same and different raw materials. When raw materials arrive at the company, they must be stored in warehouse locations with not reserved capacity. It is not possible to mix different material families in the same location or exceed the maximum allowed weight for each location. The manufacturing plant adopts a job shop configuration composed of different machines. A product requires only one specific machine to be processed. Not all machines can process all products, but one product can be processed by different machines with different times and costs. Since each unit of raw material should be completely processed, some overproduction may occur in terms of the quantity of the final product in the order. If the customer does not accept this excess, a final product inventory will be created, which becomes available-to-promise for incoming customer orders proposals. To perform a specific operation, certain tools are required to operate the machine. Each tool is characterized by a specific lifetime, defined by the number of uses it can handle which in turn, are measured by the number of product units it can process from a raw material on a specific machine. Even if a tool does not reach its maximum number of units processed within its lifetime, it is discarded after used. As with raw materials, offers from tools suppliers with different costs and delivery times are considered. 3.2. Main RM process characteristics The RM process begins with the arrival of a single customer order proposal consisting of one or several order lines with quantities of different ETO/MTO products for a common delivery date. The company must provide the client with an answer regarding the order pricing and the feasibility of achieving the requested due date. For this, the RM execution mode is offline, as the ETO/MTO strategy requires a significant amount of time to perform different activities such as: analysing raw material and tool supplier offers with different costs and delivery dates, and comparing the uncommitted availability levels of raw materials (Material Available To Promise, MATP) and storage capacity at each warehouse location (Deliver To Promise, DTP) with those required to fulfil the order. Additionally, to establish reliable delivery dates and account for production costs during order pricing, the uncommitted production capacity (Capacity Available To Promise, CTP) is also considered for lot-sizing and loading decisions. As mentioned above, final product inventory will exist if the client does not accept the excess inherent to the production process, leading to a certain uncommitted quantity of final product (Available To Promise, ATP). 3.3. Main output decisions Once the RM process is executed, the main outputs or decisions obtained are. •The order price offer to the customer and whether it is possible to meet the delivery date specified in the order. •The quantity of raw materials and tools to be purchased from the different supplier offers. •The assignment of the purchased raw materials to warehouse locations and the management of their flow entering and leaving the warehouse. •The assignment of raw materials, tools and machines to the production of each final product included in the order. •The lot-sizing and loading of the ordered products onto different machines and time periods based on the uncommitted available capacity. The best output will ensure a minimum profit for the company by setting a price offer acceptable to the customer (based on historical data and market trends) with a minimum total operation cost. Indeed, minimizing purchase, storage, and production costs will allow for a higher gross margin for the company while offering a competitive price for the delivery date, thus increasing customer satisfaction. 4. MILP model for the revenue management in the order promising process To address the aforementioned problem, a MILP referred to as RMmodel, is proposed. Fig. 3 provides a general overview of the input data, the objective function, the constraints, and the output decisions Table 5 RM process characteristics: type of orders, availability levels and tool’s lifetime. Ref Type of orders Availability levels Tool’s lifetime Proposed Promised In progress MATP ATP CTP DTP Yes No Easton and Moodie (1999) X    X  X Dobson and Yano (2002) X    X  X Defregger and Kuhn (2007) X  X X X   X Spengler et al. (2007) X    X  X Ebadian et al. (2008) X   X X   X Liu and Liu (2009) X    X  X Hintsches et al. (2010) X   X X   X Martínez and Arredondo (2010) X    X  X Chaharsooghi et al. (2011) X    X  X Kalantari and Ebadian (2011) X   X X   X Li et al. (2012) X    X  X Volling et al. (2012) X    X  X Xu and Zhang (2017) X    X  X Garmdare et al. (2018) X X X   X  X Lohnert and Fischer (2019) X    X  X Abedi and Zhu (2020) X  X X X   X Percentage 100 6,25 6,25 12,5 31,25 100 0 0 100 This paper X   X X X X J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 7 Table 6 Model characteristics: objectives and decisions. Ref Obj Obj function Network Decisions One Several Max. Benefits Min. Costs Price Quantity to serve Delivery date Lotsizing Assignment Sequencing Dimensioning of capacity Tool purchase Material purchase Storage allocation Easton and Moodie (1999) XX     X     Dobson and Yano (2002) XX     X     Defregger and Kuhn (2007) XX     X   X Spengler et al. (2007) XX     X     Ebadian et al. (2008) X  X X X X X XX Liu and Liu (2009) XXXXX X     Hintsches et al. (2010) XX     X     Martínez and Arredondo (2010) XX     X     Chaharsooghi et al. (2011) XXXXX     Kalantari and Ebadian (2011) X  X X X X X XX Li et al. (2012) XXX   X X     Volling et al. (2012) XXX   X     Xu and Zhang (2017) XX     X     Garmdare et al. (2018) XXXXX X     Lohnert and Fischer (2019) XX     X X     Abedi and Zhu (2020) XX  XX X    X Percentage 100 0 87,5 12,5 43,75 6,25 31,25 18,75 100 25 12,5 0 25 0 This paper X   X X X X X   X X X J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 8 addressed in the RM-model. 4.1. Nomenclature The nomenclatura for RM-model’s formulation is organized into indices, sets, parameters (Table 8) and decision variables (Table 9). 4.2. Revenue management mathematical programming model: RM-model 4.2.1. Objective function The RM-model aims to minimise the total costs (1) in order to offer a competitive price to the customer while ensuring a minimum profit for the company. These total costs include the inventory holding costs for final products not delivered to the customer due to an excess in production (uncommitted quantities) and those delivered to customers but manufactured prior to the delivery date; holding costs for raw materials and tools; processing costs to transform the selected raw materials into finished products; setup costs for the chosen tools and item on machines; and the costs of purchasing raw materials and tools based on the corresponding supplier offers. It is noteworthy that processing costs depend not only on the final product and the resource producing it but also on the raw material used, as substitute raw materials may exist. The best purchase offers are selected for both raw materials and tools, as a combination of price, lot size, and delivery time. The tools used have their respective purchase costs and also a setup cost depending on the number of times the tool has to be changed during production. Min[Z] = ∑ i ∑ t cinvii⋅UATPIit +∑ t<fe cinvii⋅QAit⋅(fe −t) +∑ t ∑ e ∑ uϵUEe ∑ m∈UMeu cinvmm⋅UATPMeumt +∑ h ∑ t cinvhh⋅UATPHht +∑ i ∑ mϵIMi ∑ rϵRPim ∑ t cprimr⋅QPimrt +∑ i ∑ mϵIMi ∑ rϵRPim ∑ hϵHEimr ∑ t csetuimrh⋅NUimrht +∑ m ∑ o∈MOm cmom⋅QCMom +∑ h ∑ j∈HJh chjh⋅NCHjh (1) 4.2.2. Constraints •Constraints related to final products requested in the customer order proposal The total reserved quantity (QAitʹ) of final product i in time periods t’, no later than the order delivery date (fe), should be equal to the quantity requested by the customer (qit) plus an excess quantity (QEit)(2). This excess quantity arises when the processing of complete units of raw material (e.g metallic bars) is required, even though it results in more final product units than requested by customers. If the company desires no excess, an additional constraint that sets QEit to zero should be formulated (QEit =0). ∑ tʹ≤t QAitʹ= (qit +QEit) ∀i,t=fe (2) If the company adopt the policy of not leaving raw materials halfway Table 7 Model characteristics: model technique, resolution method and validation. Ref Modelling technique Resolution method Validation LP ILP MILP NLP NILP MINLP Others Exact Heuristic Metaheuristic Matheuristic Example Realistic case Real application Easton and Moodie (1999) X       X  X  Dobson and Yano (2002)   X    X   X  Defregger and Kuhn (2007)       XX  X  Spengler et al. (2007) X       X   X Ebadian et al. (2008)   X   X X X   X  Liu and Liu (2009)   X     X  X  Hintsches et al. (2010) X      X    X Martínez and Arredondo (2010) X      X   X  Chaharsooghi et al. (2011)   X     X  X  Kalantari and Ebadian (2011)   X   X X X   X  Li et al. (2012)       XX  X  Volling et al. (2012) X      X   X  Xu and Zhang (2017) X      X    X Garmdare et al. (2018)    XXX     X Lohnert and Fischer (2019)   X    X X   X  Abedi and Zhu (2020)   X    X    X Porcentaje 37,5 0 43,75 6,25 0 6,25 25 62,5 56,25 0 0 68,75 25 6,25 This paper   X    X     X J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 9 offers are considered, the greater the probability of finding one with better conditions in terms of prices, delivery times, or lot sizes. Furthermore, the total cost is lower when the possibility of using substitute raw materials is considered. The justification for this is that the decrease in raw material purchase costs outweighs the increase in processing costs when using non-ideal raw materials. In short, the scenario that is most interesting for minimizing the costs of a specific order is when the capacity level is set to surplus, the number of supply offers is high, and final products can be produced using substitute raw materials. Table 12 shows the results of the experiments from the RM perspective, considering order price, margin, and costs disaggregated by type for each scenario. As observed, these factors vary depending on the scenario, with processing and raw material purchase costs representing the highest percentage of total costs. Finally, Fig. 6 shows the impact of the fifth experimental factor, "acceptable price", which only affects the order price and profit. To illustrate this, two specific scenarios (1-FIRM, 2-LAX) have been selected. These scenarios are characterized by having two items, a low number of offers, surplus capacity, and not considering substitute raw materials. In both cases, the costs are identical. As mentioned, the order price is calculated using two methods. The red horizontal line represents the first method of calculating the order price by considering only the total costs and the minimum allowable profit defined by the company (set at an additional 40 % of the costs). The second method, represented by a blue horizontal line, depends on the value set as the "acceptable price" for each item according to historical prices for the specific customer or market information. After calculating the order price using both methods, the higher of the two order prices is chosen. Fig. 6 clearly shows that when the acceptable price is firm, the first calculation method usually prevails. On the other hand, when it is lax, the second calculation method results in a higher order price and greater profit for the company than the expected minimum. Consequently, this figure helps clarify how the order price is calculated and validates that the model is functioning as expected. 5.4. Computational efficiency Intel® Core™ i7-7500U CPU with two 2.70 GHz processor, and a 64bit operating system, was used to solve the model. The model was implemented in Pyomo 6.5.0 and solved with the Gurobi 10.0.1 solver. Microsoft Excel databases were employed to store the input data and export the values for the decision variables. Model statistics and computational efficiency per execution are presented in Table 13. As this is an optimization problem, it is crucial that resolution times are acceptable for the company. In this case, the execution times range between 2 and 210 s (Table 13), which are very short and therefore, perfectly acceptable. In all cases, the optimal solution has been reached. Regarding the experimentation factors, it has been analysed that the acceptable price does not influence either the size of the problem or the execution time, so both cases have been grouped, considering the average execution time. Additionally, while capacity does not affect the size of the problem, it does influence resolution times. As for the other factors, the size of the problem (in terms of the number of total variables and constraints) increases as more items are included in the order, more supplier offers are available, and if the possibility of producing with substitute raw materials is considered. 6. Industrial implementation of the RM model The successful implementation of the RM Model requires to be translated into a machine-readable language by means some optimization software as well as the development of a database to store, both the input data and the optimal solution of the model. In our implementation industrial case, Pyomo 6.5.0, an open-source optimization library in Python, was chosen for programming the model due to its accessibility for both, academic and industrial users. The Gurobi 10.0.1 solver was employed for the model resolution as identified to be one of the most efficient. Finally, Microsoft Excel was used as the primary data storage tool due to its widespread adoption and ease of use in industry. Alternatively, companies may decide to use other optimization software (such as GAMS, AMPL, or MATLAB), solvers (such as CPLEX, CBC, or GLPK) or database programs (such as Microsoft Access, or SQL databases). However, it is important to note that the efficiency in solving the model may vary depending on the software and solver selected. Moreover, the choice of optimization software and database should account for factors such as compatibility, ease of integration and use, and alignment with the company’s existing technological infrastructure. The RM-model is designed to run on an event-driven basis, triggered each time an order proposal is received from a customer. In order to provide a response to the customer the RM-model should be fed with specific input data provided by various company functions such as: Engineering, Procurement, Operations, Quality, Sales and Finance Departments as shown in Table 14. The engineering department is responsible for providing the technical data related to product design and its production process. The time required to obtain these inputs may vary depending on whether the products in the order are manufactured under an ETO or MTO strategy. For ETO products, additional time may be needed to define new designs (ITO) (selecting, for instance, a set of potential raw materials and machines to process them) or adapt existing ones (RTO), whereas for MTO products, pre-existing data can typically be retrieved from the company’s databases. At this point it is important to note that artificial intelligence tools should be very useful to preselect similar designs for ETO products and estimate required data for the model. The procurement department should contact raw materials and tools suppliers to gather offers, and collect details such as lot sizes, delivery times, costs, and minimum order quantities. Procurement must ensure that the offers are up-to-date and aligned with current market conditions. The operations department should provide data on available machine and warehouse capacity as well as the uncommitted quantities available to promise raw materials, tools and final products, meanwhile the quality department provide the defect percentage obtained in produced lots during the process. The sales department should provide information related to the customer’s order, such as the product ordered, the delivery date, the price history offered to the customer and market prices. In addition, the sales department should work with the finance department to ensure that the model considers the minimum margin the company intends to achieve by fulfilling the order. Finally, the financial department must provide the remaining economic data, such costs associated with the storage of raw materials, tools and final products. After the execution of the model, and in the event that a feasible date is obtained the price to be offered to the customer for the order, as well as the decisions regarding the purchases to be made from suppliers of raw materials and tools, the allocation of raw materials and tools to the manufacture of the products that make up the order, as well as the use of the different resources available. The sales department must then communicate the price to the customer. In case an agreement was achieved with the customer, the operations department must update in the company’s ERP the different levels of availability of the company’s resources (raw materials, tools, machines, warehouses and final products), ensuring that this information is updated for future model executions. In addition, if a new design has been generated to complete this order (ITO), this information must be stored by the engineering department so that it can be used for future orders exactly for the same new designed product (MTO) or for its adaptations (RTO). As regards the prerequisites in terms of data availability and reliability needed for the RM model the existence of an integrated data systems (e.g., ERP and/or MES) to ensure data consistency, accuracy, and timeliness is strongly recommended. To support the implementation J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 16 Table 12 Experimentation RM-Model results Nº exp. Nº of references per order Nº of offers for raw materials and tools Production Capacity (CTP) Substitute Raw Materials Acceptable price Order price ( € ) Total Costs ( € ) Profit ( € ) Cost of inv. Final Prod. ( € ) Cost of inv. Tool ( € ) Cost of inv. Raw Mat. ( € ) Purchase cost of Raw Mat. ( € ) Purchase cost of Tool ( € ) Processing Cost ( € ) Setup Cost ( € ) 1 2 LOW SURPLUS NO LAX 17596.64 12568.64 5028 20.46 29.44 21.6 8780 212 3420.14 85 2 2 LOW SURPLUS NO FIRM 18450.64 12568.64 5882 20.46 29.44 21.6 8780 212 3420.14 85 3 2 LOW SURPLUS YES LAX 14000.01 9911.01 4089 20.46 27.16 64.8 6000 212 3481.59 105 4 2 LOW SURPLUS YES FIRM 18450.01 9911.01 8539 20.46 27.16 64.8 6000 212 3481.59 105 5 2 LOW ADJUSTED NO LAX 18170.75 12978.75 5192 96.47 40.64 49.5 8780 412 3420.14 180 6 2 LOW ADJUSTED NO FIRM 18450.75 12978.75 5472 96.47 40.64 49.5 8780 412 3420.14 180 7 2 LOW ADJUSTED YES LAX 14412.40 10294.40 4118 96.47 40.64 83.7 6000 412 3481.59 180 8 2 LOW ADJUSTED YES FIRM 18450.40 10294.40 8156 96.47 40.64 83.7 6000 412 3481.59 180 9 2 LOW LACKING NO LAX UNFEASIBLES 10 2 LOW LACKING NO FIRM 11 2 LOW LACKING YES LAX 12 2 LOW LACKING YES FIRM 13 2 HIGH SURPLUS NO LAX 16224.68 11588.68 4636 20.46 19.08 18 7950 76 3420.14 85 14 2 HIGH SURPLUS NO FIRM 18450.68 11588.68 6862 20.46 19.08 18 7950 76 3420.14 85 15 2 HIGH SURPLUS YES LAX 14000.56 9403.56 4597 20.46 18.76 63 5440 126 3640.34 95 16 2 HIGH SURPLUS YES FIRM 18450.56 9403.56 9047 20.46 18.76 63 5440 126 3640.34 95 17 2 HIGH ADJUSTED NO LAX 16948.59 12105.59 4843 95.21 34.4 43.2 7950 382.64 3420.14 180 18 2 HIGH ADJUSTED NO FIRM 18450.59 12105.59 6345 95.21 34.4 43.2 7950 382.64 3420.14 180 19 2 HIGH ADJUSTED YES LAX 14000.02 9852.02 4148 97.75 30.6 65.7 5440 382.64 3640.34 195 20 2 HIGH ADJUSTED YES FIRM 18450.02 9852.02 8598 97.75 30.6 65.7 5440 382.64 3640.34 195 21 2 HIGH LACKING NO LAX UNFEASIBLES 22 2 HIGH LACKING NO FIRM 23 2 HIGH LACKING YES LAX 24 2 HIGH LACKING YES FIRM 25 5 LOW SURPLUS NO LAX 45397.56 32426.83 12970.73 46.01 23.8 55.6 22260 252 9564.42 225 26 5 LOW SURPLUS NO FIRM 46450 32426.83 14023.17 46.01 23.8 55.6 22260 252 9564.42 225 27 5 LOW SURPLUS YES LAX 40024.21 28588.72 11435.49 46.01 20.2 126.6 18150 252 9743.91 250 28 5 LOW SURPLUS YES FIRM 46450 28588.72 17861.28 46.01 20.2 126.6 18150 252 9743.91 250 29 5 LOW ADJUSTED NO LAX 45427.61 32447.61 12980 64.08 25.66 51.45 22260 252 9564.42 230 30 5 LOW ADJUSTED NO FIRM 46450.61 32447.61 14003 64.08 25.66 51.45 22260 252 9564.42 230 31 5 LOW ADJUSTED YES LAX 40046.42 28604.42 11442 64.08 23.38 121.05 18150 252 9743.91 250 32 5 LOW ADJUSTED YES FIRM 46450.42 28604.42 17846 64.08 23.38 121.05 18150 252 9743.91 250 33 5 LOW LACKING NO LAX UNFEASIBLES 34 5 LOW LACKING NO FIRM 35 5 LOW LACKING YES LAX 36 5 LOW LACKING YES FIRM 37 5 HIGH SURPLUS NO LAX 43082.33 30773.09 12309.24 46.01 21.16 49.5 20675 192 9564.42 225 38 5 HIGH SURPLUS NO FIRM 46450 30773.09 15676.91 46.01 21.16 49.5 20675 192 9564.42 225 39 5 HIGH SURPLUS YES LAX 39160.73 27434.73 11726 43.71 20.18 107.7 16840 252 9926.14 245 40 5 HIGH SURPLUS YES FIRM 46450.73 27434.73 19016 43.71 20.18 107.7 16840 252 9926.14 245 41 5 HIGH ADJUSTED NO LAX 43132.84 30808.84 12324 64.08 24.34 45 20675 206 9564.42 230 42 5 HIGH ADJUSTED NO FIRM 46450.84 30808.84 15642 64.08 24.34 45 20675 206 9564.42 230 43 5 HIGH ADJUSTED YES LAX 39160 27476.1 11683.90 64.08 23.38 119.25 17000 252 9767.39 250 44 5 HIGH ADJUSTED YES FIRM 46450 27476.1 18973.90 64.08 23.38 119.25 17000 252 9767.39 250 45 5 HIGH LACKING NO LAX UNFEASIBLES 46 5 HIGH LACKING NO FIRM 47 5 HIGH LACKING YES LAX 48 5 HIGH LACKING YES FIRM J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 17 of the RM model, existing processes within the company can be leveraged. For instance, integrating the model with the company’s ERP system would facilitate automated data transfer between departments, ensuring the availability of up-to-date information and minimizing manual intervention. Additionally, processes such as procurement planning, and inventory management can naturally integrate with the model. Procurement teams, for example, can use supplier relationships and regular quotation cycles to feed updated raw material and tool offers into the model. Similarly, operations teams can align the model’s outputs with their existing practices for updating capacity plans, stock availability, and resource allocation in their regular workflows. MES systems can provide very useful data of real productivity in each machine and material. Manual or fragmented data collection methods may compromise the reliability of the model, since automated data acquisition and real-time updates are desirable. Although the RM model is not specifically designed for manual production or assembly activities, it could be applied provided that reliable information on the times and costs associated with them is available. In this case, the workforce could be treated as an additional resource with limited capacity, analogously to how machines are represented in the model. However, due to the deterministic characteristic of the proposed Fig. 6. Calculation of the order price and profit obtained. Table 13 Model statistics and computational efficiency. No. of articles No. of Offers Raw Materials Subst. Capacity (CTP) Runtime (sec) Restr. Var. Total Var. Cont. Var. Integer Var. Binary 2 REF. LOW NO SURPLUS 2.403 4067 3278 48 2786 444 2 REF. LOW NO ADJUSTED 4.585 2 REF. HIGH NO SURPLUS 2.011 4603 3794 48 3146 600 2 REF. HIGH NO ADJUSTED 5.143 2 REF. LOW YES SURPLUS 5.5525 6433 5450 48 4906 496 2 REF. LOW YES ADJUSTED 8.921 2 REF. HIGH YES SURPLUS 5.124 7177 6170 48 5418 704 2 REF. HIGH YES ADJUSTED 17.731 5 REF. LOW NO SURPLUS 102.555 9645 6875 48 6017 810 5 REF. LOW NO ADJUSTED 42.427 5 REF. HIGH NO SURPLUS 51.58 10493 7697 48 6605 1044 5 REF. HIGH NO ADJUSTED 70.265 5 REF. LOW YES SURPLUS 209.055 16905 12245 48 11257 940 5 REF. LOW YES ADJUSTED 69.23 5 REF. HIGH YES SURPLUS 178.735 18273 13577 48 12225 1304 5 REF. HIGH YES ADJUSTED 174.87 J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 18 model its reliability may decrease if the manual production activities or quality checks introduce significant variability and uncertainty. For the properly application of the RM model in these situations, two options are devise: either the companies improve the standardisation of processes to reduce this variability or otherwise the model should be extended modelling certain parameters as uncertain rather than deterministic. This last option constitutes a future research line. 7. Discussion and managerial insights The variety of work in ETO/MTO companies, the complexity of customised products, and the underlying uncertainties of the markets require integrated planning and decision-making with other core processes, such as bidding and procurement, to achieve improved results throughout the value chain (Fortes et al., 2023a; Fortes et al., 2023a). From the managerial perspective, the RM-model proposed in this paper enhances the ability of companies to manage diversity and complexity of the overall value chain by supporting the decision to accept or reject customer orders in hybrid MTO/ETO environments in several ways. •The RM-model integrates various company functions, such as engineering and production as a requirement emphasized by Wikner and Rudberg (2005) and Cannas et al. (2019, 2020) but also procurement and sales, by providing a unified approach to pricing, production allocation, lot-sizing, and lead time management, while considering different availability levels and the customer requirements. This is aligned with the approach of Fortes et al. (2023a, b) that identify production planning and control as crucial aspects for ETO companies. Coordination between engineering, production and procurement as addressed in this paper is important as Bejlegaard et al. (2021) affirm, because costs and lead times depend not only on design and production operations but also on tooling and the volatile costs of raw materials. This coordination is even more relevant when costs and delivery times vary depending on the type of raw material, the supplier, and the timing of purchases. •Mitigating uncertainties: The above cited coordination among company functions supports a realistic estimation of operating costs with profitable price setting and reliable due dates. This mitigates significant engineering and production uncertainties of ETO environments becoming challenges to be faced, as highlighted by Alfnes et al. (2021), accentuated by the additional aspect of ETO products usually sharing resources with MTO and MTS products. These challenges include overengineering, variability in lead times, and complex supplier interactions. Addressing these systemic uncertainties requires robust optimization frameworks that integrate engineering, production, and procurement decisions, as our proposal. •The RM-Model supports systematic cost management by means its ability to balance raw material, machine and tooling costs for engineering that is align with the need for cost control as emphasized in Barbosa and Azevedo (2018). It also enhances profitability by mitigating the complexity of customized production planning •The RM-model shortens the time elapsed between the arrival of a customer order proposal and the company’s response increasing its reliability. Customers often pressure companies to provide a rapid response, especially when the high volatility of raw materials costs can impact the final price setting. This forces companies to quickly agree on an order price and delivery date with customers, often without a realistic estimation of its profitability and a low reliability of the committed due date. As demonstrated by the computational efficiency in the experimental results, the RM-model is and advance tool that allows for the profitability estimation to be performed in seconds, or at most, within a few minutes as pointed out by Bejlegaard et al. (2021). Also, the reliability of the customer order due dates improves since the uncommitted availability of all resources have been considered (MATP, CTP, ATP, DTP, etc.) •The RM-model allows to set prices individually to each customer increasing the profitability, taking into account historical prices and market trends. This enables certain customer segmentation in pricing, where regular and important customers can be offered lower prices and/or earlier delivery dates compared to onetime or new customers, with the goal of retaining valuable clients and build loyalty. By integrating pricing with production and supply decisions, the model ensures that customization efforts in ETO are reflected in customer pricing, addressing gaps noted in Müller-Bungart (2009) for revenue management in ETO/MTO systems. •The RM-model can support the transition of companies in the location of the 2D-CODP along the engineering and manufacturing flows in hybrid MTO/ETO environments. As previously described, the RM-model incorporates a two-dimensional view of customer order decoupling points (CODPs) for both engineering and production, as described by Wikner and Rudberg (2005), Cannas et al. (2019, 2020). It can support certain decisions that impact engineering activities at any CODP location in ETO products, even until becoming MTO products. •For MTO products, since ideally no engineering design activities are required, in this paper it is assumed that stored information can be retrieved for the company’s databases. This includes product design, potential raw materials, tools and machines to be used, and their corresponding processing costs and lead times. Other relevant information includes historical prices set to the same customer and product, as well as the minimum margin expected by the company to accept the order. In addition to this internal information, every time a customer order arrives, external inquiries are necessary to obtain competitive offers of raw materials and tools from different suppliers, due to market prices uncertainty and volatility, availability, and lead times. Other internal inquires involve to retrieving data on uncommitted available capacity of potential production machines to process the order (CTP), tools’ lifetime, warehouse capacity (DTP), and uncommitted quantities available of raw materials (MATP) and final products (ATP). Using all this input information, the RM-model is solved. The key outputs from the optimally solved model include the decision to accept or reject the order, and the selling price offered to the client, which considers the real costs incurred to performed the order (production, inventory, and supply from selected supplier offers for raw material and tool, as well as the production plan of the order). •For ETO products, the amount of design workload of new or modified products, that is, the number of engineering hours required, depends on the engineering CODP location, which may vary widely (Amaro et al., 1999; Mintzberg, 1988). This range Table 14 Data required per department. Department Data Engineering Raw material requirements (nimr), supply lead times (tsimr), processing and setup times ((tprimr,tsetuimrh) and costs (cprimr; csetuimrh), tool durability (nhuh), and tool lifetime (slimrh). Procurement Raw material and tool purchase parameters, such as raw material weights (pmm), lot sizes (qmom, qhjh), minimum (qminom) and maximum (Mwa, order quantity, delivery times (teom; tehjh), and purchase costs (cmom, chjh). Operations Productive capacity available to promise (ctprt), stock available to promise for final products (atpi0i; atpiit), raw materials (dtpeut; atpm0eum; atpmeumt), and tools (atph0h; atphht), maximum raw materials storage capacity (Mwc), and compatibility of raw materials with specific warehouse locations (yeuft). Quality Efficiency factors related with quality defects (aprimr) Sales Delivery date proposed by the customer (fe), quantity of final products ordered (qit), acceptable price (pvpacepi), Finance Minimum profit margin (bmin), inventory costs for final products (cinvi), raw materials (cinvmm), and tools (cinvhh), J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 19 includes “pure” (engineer completely new designs), “tailored” (adapting existing designs), “standardised” (combining existing options), and “none” (using the design as is). Although assessing the feasibility of available engineering capacity is outside the scope of this work, our model is particularly well-suited for Redesign-to-Order (RTO) systems although it also holds potential for specific applications in Innovate-to-Order (ITO) settings since it can reduce the time invested for some engineering decisions, as explained before. Regardless of the engineering CODP location in ETO environments, managers should provide the same inputs to the RM-model and obtain the same outputs as those described for MTO products. The difference for ETO products as regards the MTO ones, is that these inputs may be rough estimates for potential changes, and the RM-model can be executed several times for different tentative product designs, adjusting potential raw materials, tools, and machines to produce it, along with their corresponding costs and lead times, in order to evaluate the most appropriate choice. The RM-model also supports order pricing during the contract stage for ETO products, ensuring a minimum gross margin for the company by anticipating procurement and production while guaranteeing delivery dates. •As regards the manufacturing CODP, the model is very versatile due to its ability to calculate different uncommitted availability levels (MATP, CTP, DTP, ATP). This versatility allows the RM-model to be applied whether the CODP is located in procurement or finished product manufacturing activities. Due to the consideration of excess production of finished products resulting from processing constraints, the model accounts for the uncommitted availability of finished products (ATP). The ATP should also be calculated in case the company decides to manufacture certain final products with regular demand to forecast, thus transitioning from an MTO to an MTS manufacturing strategy. It is noteworthy that the model does not support a CODP located at the assembly stage (ATO), as the production process assumed in this research does not consider assembly operations. 8. Conclusions and future research lines Globalisation and market competition increase the pressure on companies to offer a wide product catalogue, including customised products, at competitive prices with short lead times and reliable due dates. This situation can oblige companies to operate under hybrid strategies such as MTO/ETO. Many of these companies face significant uncertainty not only regarding demand but also the availability and prices of raw materials and production capacity, all of which impact their overall operation costs and delivery dates. In this context, order pricing based on manufacturing and supplier costs becomes very relevant when accepting orders. This paper presents and validates a Conceptual Framework and a RM-model designed to assist managers in order promising during the revenue management process in complex hybrid MTO/ETO environments. The review of existing research highlights the scarcity of studies focused on hybrid MTO/ETO environments, as well as the absence, to the best of the authors’ knowledge, of any mathematical programming model for the RM in such companies. Therefore, the development of the CF and RM-model constitutes the two major contributions of this research. The structured analysis of existing literature based on the CF shows that the proposed MILP model is the only found that can support the RM for companies with hybrid ETO, MTO, and even MTS production strategies. Furthermore, it enables a comparison of the distinct features of our model with existing ones: the joint modelling of operational decisions related to the revenue management in MTO/ETO environments (pricing, delivery date feasibility checking based on different uncommitted availability levels, and order acceptance), along with other specific factors (substitute materials, tools lifetime, warehouse locations), constitutes the novelties of this paper. The inclusion of ATP for final products in the RM-model, due to possible excess obtained from processing entire units of raw material, makes it possible to apply this model also for MTS final products as well. All these CODP locations are reflected in the RM-model through the consideration of different uncommitted availability levels, some of which have not been previously considered (DTP and tool lifetime) and others not simultaneously considered in earlier works (MATP, ATP, CTP), thus representing another contribution of the RM-model proposed. This feature makes the optimization tool very versatile and applicable to companies with different 2D-CODP strategies, or even within the same company sharing resources for manufacturing final products under different strategies. Additionally, this characteristic makes the model suitable for reflecting the evolution of the 2D-CODP location for certain products. In addition, both proposals, the CF and the RM model, address various challenges identified in the literature: 1) the development and implementation of advanced tools to enhance process efficiency in hybrid environments; 2) reducing the gap between theory and practice by supporting order-fulfilment strategies that include different decisions such as batch size, pricing, and procurement; 3) facilitating the integration of different functions (engineering, procurement, production, and sales) by providing a realistic estimation of operating costs, profitable price setting, and reliable due dates; and 4) supporting the transition of companies in locating the 2D-CODP within the engineering and manufacturing flows in hybrid MTO/ETO environments. The RM-model was validated through a real application in a metalmechanic company. The results show that lower costs are achieved when a greater number of offers are available, substitute raw materials are considered, and there is greater uncommitted production capacity. Likewise, having access to historical and/or market information can lead to greater profits. In terms of computational efficiency, the time required for the tool to provide a solution is manageable, ranging from a few seconds to just over 3 min. Managers can use the CF to characterize their specific problem and consult those studies sharing common characteristics. Additionally, they can execute the RM-model to support decision making during the RM process. obtaining the following benefits: reduced time spent promising orders to customers; increased reliability of the delivery date agreed with the customer; optimal selection of tools and raw material supplier offers; reduction of storage costs for raw materials, tools, and final products, and increased profit through reductions in material, production, and storage costs, as well as accurate pricing personalized according to the importance of customers. Therefore, the proposed CF and RM-model contribute to achieve the aim of this research by increasing the efficiency and profitability when promising orders for MTO and/or ETO products. As future lines of action, uncertainty could be introduced into processing times or costs. This is particularly relevant for ETO/MTO products that can be processed with different materials, tools and machines where only estimates are available and also for manufacturing environments with manual production or assembly activities. In terms of experimentation, one could also explore what happens when current constraints on the model are relaxed. It would be also interesting that decision-makers could evaluate other solutions and compare them with the optimal ones, as well as to partially define the solution by fixing the value of certain decision variables, for example, by fixing the production of a batch to a specific period or the purchase of an item from a certain supplier. Another promising direction for future research is the incorporation of engineering activities with their corresponding workload into the RM-model to further enhance its applicability in ETO contexts, especially for ITO products. Finally, linking the potential of the RMmodel for optimization and its previous results with the learning capability of artificial intelligence can be a future research avenue to configurate new or adapted products and processes better and faster. J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 20 CRediT authorship contribution statement Jose V. Cavero: Writing – original draft, Visualization, Validation, Software, Investigation, Formal analysis, Data curation. M.M.E. Alemany: Writing – original draft, Validation, Supervision, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Ana Esteso: Validation, Supervision, Methodology, Investigation, Formal analysis, Data curation. Alfredo Gim´ enez: Validation, Conceptualization. Acknowledgements This research has been partially developed in the framework of the projects entitled “Industrial Production and Logistics Optimization in Industry 4.0 (i4OPT)” funded by Generalitat Valenciana (PROMETEO/ 2021/065) and “DiMAT–Digital Modelling and Simulation for Design, Processing and Manufacturing of Advanced Materials”, funded by the European Union’s Horizon Europe research and innovation programme under the Grant Agreement no.101091496. The authors acknowledge the support of the “Factor Ingeniería y Decoletaje, S.L” company for the data provided and the model validation. Data availability I have shared the link to my data https://doi. org/10.5281/zenodo.10887688 References Abedi, A., Zhu, W., 2020. An advanced order acceptance model for hybrid production strategy. J. Manuf. Syst. 55, 82–93. Alarc´ on, F., Alemany, M., Ortiz, ´ A., 2007. El Proceso de Comprometer Pedidos (Order Promising) y su relaci´ on con la Planificaci´ on de la Producci´ on. XI Congreso de Ingeniería de Organizaci´ on 335–345. https://doi.org/10.1016/j.jmsy.2020.02.012. Alemany, M.M.E., Alarc´ on, F., Ortiz, A., Lario, F.C., 2008. Order promising process for extended collaborative selling chain. Prod. Plann. Control 19, 105–131. https://doi. org/10.1080/09537280801896011. Alfnes, E., Gosling, J., Naim, M., Dreyer, H.C., 2021. Exploring systemic factors creating uncertainty in complex engineer-to-order supply chains: case studies from Norwegian shipbuilding first tier suppliers. Int. J. Prod. Econ. 240, 108211. https:// doi.org/10.1016/j.ijpe.2021.108211. Alfnes, E., Gosling, J., Naim, M., Dreyer, H.C., 2023. Rearticulating supply chain design and operation principles to mitigate uncertainty in the Norwegian engineer-to-order shipbuilding sector. Int. J. Prod. Econ. 262, 108903. https://doi.org/10.1016/j. ijpe.2023.108903. Amaro, G., Hendry, L., Kingsman, B., 1999. Competitive advantage, customisation and a new taxonomy for non make-to-stock companies. Int. J. Oper. Prod. Manag. https:// doi.org/10.1108/01443579910254213. Barbosa, C., Azevedo, A., 2018. Hybrid modelling of MTO/ETO manufacturing environments for performance assessment. Int. J. Prod. Res. https://doi.org/ 10.1080/00207543.2017.1421788. Bejlegaard, M., Sarivan, I.-M., Waehrens, B.V., 2021. The influence of digital technologies on supply chain coordination strategies. J. Global Oper. Strat. Sourc. 14 (4), 636–658. https://doi.org/10.1108/jgoss-11-2019-0063. Cannas, V.G., Gosling, J., 2021. A decade of engineering-to-order (2010–2020): progress and emerging themes. Int. J. Prod. Econ. 241, 108274. https://doi.org/10.1016/j. ijpe.2021.108274. Cannas, V.G., Gosling, J., Pero, M., Rossi, T., 2019. Engineering and production decoupling configurations: an empirical study in the machinery industry. Int. J. Prod. Econ. 216, 173–189. https://doi.org/10.1016/j.ijpe.2019.04.025. Cannas, V.G., Gosling, J., Pero, M., Rossi, T., 2020. Determinants for order-fulfilment strategies in engineer-to-order companies: insights from the machinery industry. Int. J. Prod. Econ. 228, 107743. https://doi.org/10.1016/j.ijpe.2020.107743. Chaharsooghi, S.K., Honarvar, M., Modarres, M., 2011. A multi-stage stochastic programming model for dynamic pricing and lead time decisions in multi-class make-to-order firm. Sci. Iran. 18, 711–721. https://doi.org/10.1016/j. scient.2011.05.018, 2011. Chen, J.-M., Hsu, Y.-T., 2017. Revenue management for durable goods using trade-ins with certified pre-owned options. Int. J. Prod. Econ. 186, 55–70. https://doi.org/ 10.1016/j.ijpe.2017.01.019. Defregger, F., Kuhn, H., 2007. Revenue management for a make-to-order company with limited inventory capacity. OR Spectrum 29, 137–156. https://doi.org/10.1007/ s00291-005-0016-1. Dobson, G., Yano, C.A., 2002. Product offering, pricing, and make-to-stock/make-toorder decisions with shared capacity. Prod. Oper. Manag. 11, 293–312. https://doi. org/10.1111/j.1937-5956.2002.tb00188.x. Easton, F.F., Moodie, D.R., 1999. Pricing and lead time decisions for make-to-order firms with contingent orders. Eur. J. Oper. Res. 116, 305–318. Ebadian, M., Rabbani, M., Jolai, F., Torabi, S.A., Tavakkoli-Moghaddam, R., 2008. A new decision-making structure for the order entry stage in make-to-order environments. Int. J. Prod. Econ. 111, 351–367. https://doi.org/10.1016/j.ijpe.2007.01.004. Fortes, C.S., Tenera, A.B., Cunha, P.F., 2023a. Engineer-to-Order challenges and issues: a systematic literature review of the manufacturing industry. Proc. Comput. Sci. 219, 1727–1734. https://doi.org/10.1016/j.procs.2023.01.467. Fortes, C.S., Tenera, A.B., Cunha, P.F., Teixeira, J.P., 2023b. Engineering-to-order manufacturing: a criticality analysis of key challenges and solutions based on literature review. Adv. Prod. Eng. Manag. 18 (2), 187–198. https://doi.org/ 10.14743/apem2023.2.466. Garmdare, H.S., Lotfi, M.M., Honarvar, M., 2018. Integrated model for pricing, delivery time setting, and scheduling in make-to-order environments. J. Ind. Eng. Int. 14, 55–64. https://doi.org/10.1007/s40092-017-0205-y. Grillo-Espinoza, H., Alemany, M.M.E., Ortiz, A., 2016. A review of mathematical models for supporting the order promising process under lack of homogeneity in product and other sources of uncertainty. Comput. Ind. Eng. 91, 239–261. https://doi.org/ 10.1016/j.cie.2015.11.013. Grillo-Espinoza, H., Alemany, M.M.E., Ortiz, A., De Baets, B., 2019. Possibilistic compositions and state functions: application to the order promising process for perishables. Int. J. Prod. Res. 22 (57), 7006–7031. https://doi.org/10.1080/ 00207543.2019.1574039. Harris, F.H. de B., Pinder, J.P., 1995. A revenue management approach to demand management and order booking in assemble-to-order manufacturing. J. Oper. Manag. 13, 299–309. https://doi.org/10.1016/0272-6963(95)00029-1. Hintsches, A., Spengler, T.S., Volling, T., Wittek, K., Priegnitz, G., 2010. Revenue management in make-to-order manufacturing: case study of capacity control at Thyssen Krupp VDM. Bus. Res. 3, 173–190. https://doi.org/10.1007/BF0334272. Kalantari, Rabbani M., Ebadian, M., 2011. A decision support system for order acceptance/rejection in hybrid MTS/MTO production systems. Appl. Math. Model. 35 (3), 1363–1377. https://doi.org/10.1016/j.apm.2010.09.015. Kimes, Sheryl E., 1989. Yield management: a tool for capacity-constrained service firms. J. Oper. Manag. 8 (4), 348–363. Li, X., Wang, J., Sawhney, R., 2012. Reinforcement learning for joint pricing, lead-time and scheduling decisions in make-to-order systems. Eur. J. Oper. Res. 221, 99–109. Liu, S.C., Liu, C.H., 2009. Simultaneous pricing, due-date setting and scheduling for make-to-order firms with contingent orders. J. Chin. Inst. Ind. Eng. 26, 32–43. https://doi.org/10.1080/10170660909509119. Lohnert, N., Fischer, K., 2019. Booking limit based revenue management approaches for customer-value oriented make-to-order production. In: Lecture Notes in Logistics. Springer Science and Business Media B.V, pp. 268–282. https://doi.org/10.1007/ 978-3-030-29821-0_18, 2019. Lorente-Leyva, L.L., Alemany, M.M.E., Peluffo-Ord´ o˜ nez, D.H., 2024. A conceptual framework for the operations planning of the textile supply chains: insights for sustainable and smart planning in uncertain and dynamic contexts. Comput. Ind. Eng. 187, 109824. https://doi.org/10.1016/j.cie.2023.109824. ISSN 0360-8352. Martínez, E., Arredondo, F., 2010. Order acceptance for revenue management and capacity allocation in make-to-order batch plants. Comput. Aided Chem. Eng. 28 (2010), 1189–1194. Mintzberg, H., 1988. Generic strategies: toward a comprehensive framework. Adv. Strat. Manag. 5, 1–67. Müller-Bungart, M., 2007. Revenue management with flexible products: models and methods for the broadcasting industry. In: Lecture Notes in Economics and Mathematical Systems. Springer, Berlin, pp. 1–313. Ould, M.A., 1995. Business Processes: Modelling and Analysis for Re-Engineering and Improvement. New York. Raza, S.A., Rathinam, S., Turiac, M., Kerbache, L., 2018. An integrated revenue management framework for a firm’s greening, pricing and inventory decisions. Int. J. Prod. Econ. 195, 373–390. https://doi.org/10.1016/j.ijpe.2016.11.014. Rehkopf, S., Spengler, T., 2005. Revenue Management in a make-to-order Environment, pp. 470–478. https://doi.org/10.1007/3-540-27679-3_58. Seuring, S., Müller, M., 2008. From a literature review to a conceptual framework for sustainable supply chain management. J. Clean. Prod. 16, 1699–1710. https://doi. org/10.1016/J.JCLEPRO.2008.04.020. Spengler, T., Rehkopf, S., Volling, T., 2007. Revenue management in make-to-order manufacturing-an application to the iron and steel industry. OR Spectrum 29, 157–171. https://doi.org/10.1007/s00291-005-0024-1. Talluri, K., van Ryzin, G., 2004. The Theory and Practice of Revenue Management. Springer, New York. Volling, T., Akyol, D., Wittek, K., Spengler, T., 2012. A two-stage bid-price control for make-to-order revenue management. Comput. Oper. Res. 39, 1021–1032. Wikner, J., Rudberg, M., 2005. Integrating production and engineering perspectives on the customer order decoupling point. Int. J. Oper. Prod. Manag. 25 (7), 623–641. https://doi.org/10.1108/01443570510605072. Xu, X., Zhang, P. He, Xu, X., 2017. Production and pricing problems in make-to-order supply chain with cap-and-trade regulation. Omega (Westport) 66, 248–257. Zatta, D., 2016. Revenue Management in Manufacturing. State of the Art, Application and Profit Impact in the Process Industry. Springer. J.V. Cavero et al. International Journal of Production Economics 288 (2025) 109715 21