Procurement for supply chain integration: responding to complexity in engineer-to-order defence projects

European defence supply chains, weakened by decades of budget reductions, have become increasingly fragile. The sector's rapid capacity expansion has now been exacerbated by the war in Europe and Geopolitical instability. Ambitious proposed defence investments in complex technological equipment are exposing the vulnerabilities of globalised supply chains (The Economist, 2023). To manage these challenges, governments are pursuing public procurement reforms to tackle cost overruns and delays through closer industry ties and improved integration (Pfeifer and Fisher, 2024). For the defence industry, traditional procurement regimes, designed for stable schedules and long-term supplier relationships, are increasingly misaligned with today's compressed timelines, increasing defence demand and complex global supply chains. This misalignment foregrounds the central issue of how procurement, at the project level (cf. Moretto et al., 2022), shapes responses through integration practices, both internally and externally, when faced with complexity.

SCI is the degree to which a manufacturer strategically collaborates with supply chain partners and manages processes within and across organisations (Flynn et al., 2010). Prior work links SCI to resilience, efficiency and financial outcomes through improved information sharing, coordinated decision making, adaptability and process optimisation (Agyei-Owusu et al., 2022; Ataseven and Nair, 2017; Flynn et al., 2016; Frohlich and Westbrook, 2001), while emphasising that effects are contingent on operations strategy and competitive priorities (Leuschner et al., 2013). For this study, we characterise defence projects as engineer-to-order (ETO) situations, where ETO denotes project-based development and delivery of customised products to specific customer requirements, often with limited standardisation and late design freeze (Alfnes et al., 2021; Cannas and Gosling, 2021). Such projects require significant efforts to manage complexity and its different forms (Maylor and Turner, 2017). Studies show that ETO supply chains typically combine high levels of “structural complexity”, including multi-tier networks and bespoke components, with “dynamic complexity”, including shifting customer requirements and uncertain timelines (Birkie and Trucco, 2016; Bozarth et al., 2009). Established integration approaches from repetitive manufacturing, therefore, translate only partially to ETO conditions, and evidence on SCI remains skewed toward high-volume contexts (Khanuja and Jain, 2019).

A key premise for this work is that procurement is strategically positioned to address these contextual issues, as well as the theoretical misalignment of SCI in ETO. Procurement sits at the intersection of SCI dimensions and mediates across technical, organisational and market domains (Laari et al., 2023; Picaud-Bello et al., 2022; Scherer and Biemans, 2025). In addition, key procurement activities such as supplier relationship management, early supplier involvement, sourcing choices, timing of component deliveries, and coordination routines directly shape lead times and system integration in ETO projects (Gosling and Naim, 2009; Hicks et al., 2000; Mello and Strandhagen, 2011; Reitsma et al., 2024). Governance choices that balance contractual and relational mechanisms, including moves toward procuring complex performance, are likewise enacted through procurement to tackle systemic complexity (cf. Roehrich and Lewis, 2014). Defence projects heighten these issues: large, technologically advanced systems are managed over long-life cycles and often orchestrated by prime contractors within oligopolistic structures (Caldwell and Howard, 2014; Howard et al., 2016; Johnsen et al., 2009). What remains underexplored is the way in which procurement can respond to structural and dynamic complexity to facilitate SCI in ETO defence projects. Therefore, the overarching purpose is to explore procurement responses to varying levels of complexity.

The research is guided by the following research question: “How can procurement respond to the integration of ETO supply chains across different types and levels of complexity?” The study is based on a qualitative embedded case design, incorporating interviews with procurement employees, project leaders, engineers and a customer representative. The design allows us to identify SCI factors salient to ETO defence supply chains and to examine procurement responses under different combinations of structural and dynamic complexity, culminating in a framework that links complexity profiles to integration approaches.

This study addresses a timely misalignment of resources, capacity and practices in the context of ambitious expansion of investment in the defence industry. We observe a mismatch where SCI literature is mostly skewed towards high-volume make-to-stock type industries, and there is not enough clarity on procurement's mediating strategic role. We contribute theoretically by showing how procurement addresses varying complexities through the integration of complexity theory and procurement literature. By distinguishing structural and dynamic dimensions of complexity, it develops a framework for tailoring procurement responses to balance integration efforts. The findings emphasise procurement's ongoing mediating role by underscoring procurement's strategic influence on SCI in ETO.

Cannas et al. (2020) define ETO as an order fulfilment strategy where engineering and production flows are driven almost entirely by the end customer. The order-fulfilment strategy refers to the way a supply chain responds to a customer order, encompassing the flow of activities from sales inquiry to product delivery (Cannas et al., 2020). Key structural activities to manage ETO order fulfilment flow include quotations, technical and commercial development, design, production, delivery, commissioning and after-sales (Adrodegari et al., 2015). An important distinction for this work is that many, if not all, of these activities will require procurement teams to procure goods and services for customer-specific needs across the order fulfilment flow. Several typologies have been developed for ETO-type firms, which help convey their key characteristics. Hicks et al. (2001) emphasize complexity of product depth and Job Shop environment, while Willner et al. (2016) focus on low volumes and high number of engineering hours needed per order, when compared with non-ETO type firms.

Transferring supply chain practices from high-volume manufacturing to engineering-intensive ETO sectors is problematic due to irregular order patterns, low procurement volumes, and unique customer specifications that require frequent supplier adaptation (cf. Gosling et al., 2020). Unlike manufacturers with consistent, high-volume procurement, ETO firms purchase components in low volumes and at irregular intervals, which limits their influence over suppliers (Cannas and Gosling, 2021; Hicks et al., 2000). Despite these challenges, SCI is critical across tendering, product development, and production stages to ensure ETO products meet unique customer specifications efficiently (Cannas et al., 2019; Gosling and Naim, 2009). Indeed, integration facilitates the fulfilment of customer expectations while minimising costly errors and delays (Mello et al., 2015b).

In ETO projects, volatile specifications and exposure to engineering changes drive redesign and rework, which increase costs and lead-times (Hicks et al., 2001; Iakymenko et al., 2022; Shurrab et al., 2022). Overlapping phases of design, procurement and manufacturing, combined with early commitment to delivery dates (Bhalla et al., 2022; Hicks et al., 2001; McGovern et al., 1999), create challenges for design specifications, production planning, and capacity control (Fortes et al., 2023; Mello et al., 2015a). Low and irregular purchasing volumes limit supplier leverage (Cannas and Gosling, 2021; Hicks et al., 2000), and intermittent supplier involvement causes information gaps (Bäckstrand and Fredriksson, 2022; Cohee et al., 2018). For long-life systems and products, lifecycle and obsolescence risks need early attention (Howard and Miemczyk, 2010; Romero Rojo et al., 2010). Design reuse and component standardisation reduce costs and lead time, reinforcing procurement-led supply chain integration across tendering, development and execution (Hicks et al., 2001; Strandhagen et al., 2018).

Integration in ETO-supply chains involves managing issues such as complex supplier relationships, negotiating with limited bargaining power due to low-volume orders, and balancing design specifications between functionality, additional features and cost (Hicks et al., 2000). ETO-companies with limited leverage can address these constraints by establishing strong supplier relationships, communicating the value of long-term collaboration, embracing flexibility, and aligning procurement strategies with business goals (Christiansen and Maltz, 2002). However, poor information sharing can undermine coordination, particularly when suppliers are not continuously involved in the process (Bäckstrand and Fredriksson, 2022; Kjersem and Giskeødegård, 2020). Additionally, early customer engagement improves specification accuracy and execution, whereas late-stage involvement and changes increase project risk (Dixit et al., 2019). Customer-specific requirements, order changes, and varied product knowledge contribute to supply chain uncertainty (Shurrab et al., 2022), heightening the need for strategic procurement practices. Collectively, these issues stem from the inherent complexity of ETO environments.

Particularly, procurement integration is crucial, as up to 80% of materials and components originate from suppliers (Cigolini et al., 2022). Moretto et al. (2022) highlight the importance of aligning procurement with both internal (time pressures, resource dedication) and external (customer involvement, project uniqueness) factors. They identify two procurement archetypes: procurement-focused (centralised, efficiency-driven) and project-focused (decentralised, project-specific), with a hybrid model in between. Reitsma et al. (2023) underline the importance of integrating procurement with R&D early in the design phase, assigning appropriate representatives, and clarifying roles.

Recent studies frame procurement as a strategic mediator, acting as a knowledge broker that translates customer needs, supplier capabilities, and internal objectives to facilitate integration across boundaries. Procurement supports this through coordinating evolving requirements, resolving demand-supply imbalances (Laari et al., 2023), and leveraging supplier knowledge in product development (Picaud-Bello et al., 2022). In ETO, procurement orchestrates integration from early development stages by aligning technical designs with supplier inputs (Eriksson, 2015; Reitsma et al., 2024). This study outlines procurement as a strategic, mediating mechanism that enables alignment across internal functions and external partners under conditions of uncertainty.

In ETO-projects, design and procurement often begin before contract finalisation, demanding early and sustained integration in tendering, development, and supplier engagement (Bhalla et al., 2022). Procuring long-lead-time items ahead of design finalisation can shorten timelines, but it also exposes projects to the risk of late-stage changes (McGovern et al., 1999). Strategic sourcing choices, especially between standard and custom components, hinge on timing and design certainty (Strandhagen et al., 2018). During tendering, procurement estimates costs and lead times while selecting suppliers who influence project outcomes (Buzzetto et al., 2020). To perform effectively, procurement must maintain tight coordination across design, R&D, and operations, informed by supplier and market insights (Guy and Dale, 1993; Reitsma et al., 2023).

Clear, technically robust specifications are essential, especially when developed collaboratively between design and procurement teams (Dowlatshahi, 1992). Functional flexibility is valuable (Hicks et al., 2001), but technical clarity ensures manufacturability and facilitates supplier contributions. Early supplier involvement improves performance and reduces complexity when supported by formal agreements and shared objectives (Cohee et al., 2018). Nonetheless, poor internal alignment can strain supplier relationships and reduce responsiveness (Ellegaard and Koch, 2014; Foerstl et al., 2013), underlining the importance of procurement-led integration across functions.

Procurement also navigates trade-offs between cost, availability, and technical quality, often in collaboration with design teams (Dowlatshahi, 1992). Early awareness of engineering changes minimises rework and post-implementation inefficiencies (Iakymenko et al., 2022), indicating the importance of early customer engagement (Dixit et al., 2019). Long operational lifespans in defence systems introduce further complexity, particularly in sustaining in-use performance amid evolving threats and shifting technical landscapes (Howard and Miemczyk, 2010). Addressing obsolescence proactively from the design phase is critical, especially where ageing mechanical or electronic components may be unavailable in later stages (Romero Rojo et al., 2010). Procurement thus supports lifecycle adaptability, ensuring systems remain operable and supportable over time. Thus, procurement is not only a transactional function but a strategic enabler of integration in complex ETO environments.

Complexity, characterised by uncertainty, structural interdependencies, and dynamic changes, is central to ETO-projects due to their high customisation and novelty (Maylor and Turner, 2017). Complexity, we define as the variety of elements, i.e., products, systems, and organisational entities, and their interactions within a system (Choi and Krause, 2006). Influenced by Bozarth et al. (2009), we classify complexity into structural and dynamic dimensions, each stemming from internal (within the firm) or external (across the supply/value chain) sources (Baccarini, 1996; Williams, 1999). Baccarini (1996) conceptualises project complexity in terms of structural differentiation, meaning the number of elements and components, and the interdependency between them. This approach captures ETO's advanced technical architectures and supplier and customer network links (Choi and Krause, 2006). Dynamic complexity reflects uncertainty in project goals and methods, evolving from specifications and frequent changes typical in ETO settings (Tatikonda and Rosenthal, 2000; Williams, 1999). Together, these two dimensions cover internal and external complexity (see Table 1), aligning with ETO-projects’ unique products, evolving requirements, and demands for integration in the supply chain (Baccarini, 1996; Choi and Krause, 2006; Tatikonda and Rosenthal, 2000; Williams, 1999).

Complexity in ETO environments is inevitable and requires active management. As Maylor and Turner (2017) argue, it may be reduced but never fully eliminated, necessitating proactive rather than reactive approaches. In line with their “understand–reduce–respond” framework, we suggest that procurement has a role to play in managing contingency responses to complexity. SCI involves internal, supplier, and customer dimensions. Internal integration is increasingly seen as a prerequisite for effective external integration (Ataseven and Nair, 2017; Flynn et al., 2010; Munir et al., 2020). Yet, despite the acknowledged importance of internal integration, procurement's role in orchestrating internal–external linkages remains underdeveloped in SCI theorising.

Despite growing interest in procurement's role in SCI within ETO contexts, essential aspects of its theoretical foundations remain underexplored. First, most SCI frameworks are developed for stable high-volume manufacturing and do not account for the complexity and uncertainty inherent in ETO environments. Second, while procurement is acknowledged as important, its role as a strategic mediator in the nexus of internal, supplier, and customer integration under ETO complexity remains undertheorized. Third, there is a limited understanding of how different types of complexity, structural and dynamic, shape procurement responses that enable SCI.

This study emerges from a research program examining procurement in defence ETO-projects, a critical sector innovating products amid contemporary military conflicts (Dixit et al., 2019). Employing a multiple-case approach to investigate complex phenomena in real-world settings (Eisenhardt, 1989), the study uses a priori constructs, including SCI dimensions, to illuminate how procurement activities shape downstream supply chain decisions from both upstream and internal perspectives. Established theories introduced into a less-explored context (DuHadway et al., 2022) uncover unique dynamics and extend existing frameworks. Drawing on case research principles in operations management (Voss et al., 2002), the methodology integrates multiple literature streams: ETO SCM, SCI, and procurement with empirical data from interviews, secondary sources, triangulation, and case descriptions.

The findings focus on DefCo (pseudonym), a large Swedish multinational company engaged in complex ETO-projects. Exploratory research, valuable for uncovering unexpected phenomena to inform theory-building (Eisenhardt, 1989), was chosen to address the limited studies on procurement integration in ETO. This approach provides academic insights and managerial implications. While the findings are grounded in a specific industry context, they may apply to related ETO contexts. The study's focus enables an in-depth exploration of unique aspects while addressing the dynamics and challenges faced by ETO-companies.

The Defence Procurement Agency (DPA) procures, develops, and delivers equipment and services to the armed forces and military units (Ekström et al., 2021). In the simplified supply chain illustrated in Figure 1, DefCo acts as a systems integrator, sourcing components and technology from suppliers to assemble final products ordered by the DPA. These products are delivered directly or through the DPA to the armed forces and end-users. DefCo's complex and lengthy development projects require close collaboration with suppliers to gather price estimates and technical information, as well as with the DPA to align product requirements. This process enables DefCo to estimate project costs and provide accurate quotations.

The projects were selected based on the following: (1) each had to be an ETO-project with at least a hybrid-focused procurement organisation (Moretto et al., 2022); (2) a project-team needed to be available for interviews focusing on procurement issues; (3) preference was given to projects spanning several years; and (4) the projects should exhibit varied complexity, including the number of suppliers, supplier site development, customer involvement, concurrent development, and project lead times.

The sample includes 26 interviews from three projects (1,538 min total) covering different roles (Table 2); primarily of procurement professionals, complemented by project leaders who provided a holistic view of the projects and explained how procurement is integrated internally and externally. Respondents from the design department were included to capture internal integration insights. Initial interviews with the head of procurement highlighted that collaboration with the design department was crucial for fostering long-term, proactive integration within and across projects. Additional interviews with operational excellence professionals, who acted as a bridge between engineering and procurement, shed light on managing costs and enhancing collaboration. A customer interview enriched the findings by providing insights into downstream integration. The selection criteria emphasised respondents' qualifications and experience, ensuring a comprehensive understanding of the projects and integration processes. This allowed respondents to relate their integration practices to past experiences in various industries and projects. This contributed to validity, as their insights helped ensure that the study accurately captured the studied phenomenon (cf. Golafshani, 2003; Voss et al., 2002). Several interviews were conducted by multiple authors, promoting diverse perspectives and strengthening confidence in the findings through consistent observations (Eisenhardt, 1989). We triangulated our findings using confidential internal documents (process and role descriptions, project presentations) alongside interview data. Respondents and a designated manager reviewed all transcripts for clarity, accuracy, and confidentiality. Follow-up interviews then resolved any discrepancies and clarified outstanding issues.

Online interviews were conducted, recorded, and transcribed verbatim. The transcripts were coded using Scrivener software, following a structured scheme that moved iteratively from first-order concepts (informant accounts) to second-order themes (integration dimensions) and then to aggregate SCI dimensions. Our analysis followed an abductive approach (Dubois and Gadde, 2002), moving iteratively between empirical data and existing SCI constructs. Specifically, while coding began at the level of informant accounts, we used the SCI dimensions as a sensitizing framework to guide how concepts were grouped and interpreted. In this sense, our approach reflects an adaptation of the Gioia methodology (Gioia et al., 2013): we preserved the progression from informant terms to higher-order categories but did so within a theory-informed structure, allowing inductive insights to refine and extend the SCI-based coding frame. To evaluate internal and external structural and dynamic complexity (outlined in Table 4), procurement responses were coded on a high/medium/low scale based on interview data (bracketed in the table). The results were then compared across the three cases (Alpha, Beta, Gamma), enabling us to identify recurring patterns in procurement's mediating roles. Coding consensus was achieved through peer debriefing (e.g., Tables 3 and 4), consistent with O'Connor and Joffe (2020), who emphasise negotiated agreement as a reflexive means to ensure coding reliability and transparency. This negotiated consensus formed the basis for mapping the cases onto the complexity integration framework (Figure 2), illustrating how procurement's responses vary across different combinations of structural and dynamic complexity. Throughout the process, we remained reflexive, drawing on both insider and outsider perspectives to balance contextual depth and critical distance.

This section presents within-case findings for Projects Alpha, Beta and Gamma. Then it turns to cross-case themes organised around the SCI-dimensions, maps them to the integration dimensions (2nd-order-themes) and responses to structural and dynamic complexity (1st-order-concepts), as reflected in Table 3.

Project Alpha originated in early studies decades ago, with development and production starting about ten years ago ^[1]. It exhibits overlapping design, development and production phases, creating high dynamic complexity. Initially, knowledge gaps were addressed through intensive early integration with the customer. Because no comparable product had been built for decades, a near-complete restart of the supply chain was required, which comprised hundreds of sub-systems, making the technical scope structurally complex. The upstream supply chain mixes COTS-components with systems produced to DefCo drawings across a range of TRLs (from basic research to proven technology), and supplier systems undergo factory acceptance tests before delivery. The supplier base comprises approximately 1,000 suppliers, with several single-source items that limit leverage and increase risk.

Internal integration has become more formalised over time. Alpha embeds a procurement representative in cross-functional teams ^[2] and uses a group of procurement coordination managers to handle work coordination, supplier management, material planning, subcontracting and budget follow-up. Cross-functional sourcing is used to support prototype sourcing and make-or-buy choices, with coordinated reviews of early design options to inform cost and technology selection. Alpha applied design-to-cost. Later, some low-price selections increased total cost because new suppliers struggled to interpret demanding requirements. Alpha combines high structural and high dynamic complexity due to many interdependent systems and evolving requirements.

Supplier integration emphasises early involvement of ETO-experienced partners and balancing mature TRLs with areas needing lower-TRL development. Flexibility is key, as Alpha must align multiple interdependent systems progressing at different speeds. As one respondent noted:

Integration is early and intensive for long-lead-time and selected major systems. Strong relationships and close follow-up are prioritised to ensure quality and lead times, despite the resource burden:

Customer integration is exceptionally close. A mirrored organisation for system engineers and daily on-site customer personnel enabled joint decision-making on evolving technologies. Procurement and the customer collaborated weekly at the sub-system level. This iterative involvement helped manage obsolescence and technology change over long timelines.

Post-delivery, Alpha highlights operational readiness via proactive planning. Although current contracts omit maintenance, older products are covered by framework agreements for configuration management and annual service. Obsolescence management adds dynamic complexity: suppliers are expected to signal EOL, but notifications can be inconsistent.

Project Beta started around the same time as Alpha and was developed for two customers who required slightly different end products. The project draws on roughly 100 COTS suppliers plus 20 drawing-based suppliers, combining proven items with developed solutions. Procurement was involved later in the process, indicating potential benefits from earlier cross-functional coordination with design. As one procurement manager noted:

Internally, one coordinator mediates between the project and procurement ^[3]. Beta used design-to-cost; a product manager taught new design engineers, using the Kraljic matrix, how early decisions influence costs and sourcing flexibility ^[4]; aiming to avoid single-source components. As costs later rose, the procurement team challenged existing suppliers to meet new targets, illustrating the need to embed cost thinking from the concept phase to mitigate dynamic complexity.

Supplier integration is staged. Beta sought high-TRL solutions to reduce risk but sometimes adopted lower-TRL technologies to meet specific customer demands. Where cost pressures intensified, supplier strategies were revisited, including developing alternatives or switching suppliers (a response to structural complexity). Procurement prioritised lead time over price, with intensive follow-up to protect quality and delivery.

Customer integration followed a milestone-based pattern. In some systems, customers pre-defined certain suppliers; otherwise, procurement received customer needs via internal channels after contract award. The customers engaged in reviews, validation and trade-off decisions throughout development.

Gamma has been in development for nearly a decade ^[5], with ongoing component acquisition and an upcoming verification phase to meet standards before production. Production will be outsourced to a partner due to previous export requirements for local manufacturing, as some parts are manufactured in Sweden, and final assembly is conducted by the partner, adding to the structural complexity. Deliveries will be ongoing and in greater volumes than Alpha. The supplier base is small, around 20 suppliers, but it includes single-source items. In one instance, a single-source supplier prioritised a higher-volume customer, delaying Gamma and highlighting the dynamic complexity in capacity allocation.

Internally, Gamma uses a single mediator to link project and procurement:

Rather than design-to-cost, the programme prioritised time and functionality, and production teams were given stronger authority over cost-related decisions to align design and manufacturability better. Although a procurement manager noted that an improvement would be to have a checkpoint already during the design phase.

Early supplier involvement was crucial for managing long lead-time items, and procurement coordinated closely with engineering to secure components ahead of verification. Although the customer had no direct role in sourcing decisions, they influenced component selection for safety-critical systems. Procurement navigated these inputs while maintaining supplier alignment and flexibility. Given the small scale and custom nature of the project, securing supplier commitment required proactive relationship management and an understanding of ETO-specific demands. The procurement team focused less on cost and more on lead time, quality, and adaptability, key elements in managing dynamic complexity in Gamma.

Customer integration is managed through integrated project teams that keep the customer informed of progress. The customer did not participate directly in sourcing decisions, although they influenced component choices where safety implications were significant. Deliveries will be ongoing, with Gamma produced in greater volumes than Alpha.

All three projects adapted to complex procurement environments in different ways. Alpha managed extensive structural complexity, requiring intensive integration with customers and suppliers. Beta balanced dual-customer needs with cost control. Gamma prioritised time and functionality over cost, relying on selective customer input and a smaller supplier base. These differences shaped distinct procurement logics across the SCI dimensions. The analysis below is organised around the integration dimensions in Table 3.

Procurement mediation roles differ slightly. Alpha employs multiple project mediators to coordinate sub-projects, align material flows, and link with engineering and supplier quality – a routine-driven model. Beta and Gamma instead use a single point-of-contact connecting buyers to commodity and programme leads – broader but lighter. All three emphasize hybrid profiles and rotation to embed engineering and project skills in procurement. Cross-functional collaboration is central: Alpha promotes integration through procurement teams and tollgate governance; Beta via category teams aligning technology roadmaps; and Gamma through joint meetings and shared plans following early fragmentation. Cost-consciousness design takes distinct forms. Alpha's design-to-cost approach highlights that the cheapest option can raise total costs through supplier and internal inefficiencies. Beta embeds cost thinking in design reviews and uses Kraljic logic to prevent single-sourcing. Gamma prioritizes functionality and speed. In managing obsolescence, Alpha is most process-oriented, front-loading visibility on long-lead items and mediating between integrated logistics, procurement, and engineering to handle lifecycle guarantees. Beta addresses obsolescence episodically through last-time buys during delays.

Early supplier involvement is central but context-specific. Alpha engages suppliers on critical or immature technologies and long-lead systems, sometimes sole-source, to co-define requirements and risks before standard procurement. Beta typically involves supplier's post-customer award, focusing on prototypes and lead-time compression, while Gamma engages them in bids and early development to assess capabilities and requirement impacts, vital for low-volume custom work. Supplier capabilities alignment diverge: Alpha matches high-TRL predictability with ETO-experienced partners, Beta redesigns out and replaces underperformers, and Gamma favours partners designing to functional specs with transparent system impact. Supplier lock-in risks exist across cases: Alpha faces it when only one supplier can deliver required parts, mitigating through EOL agreements and obsolescence management; Beta's design specified a patented component, which restricted supplier choice even though functional alternatives existed; Gamma relied on single source-suppliers, making strong contracts vital as switching at critical stages proved slow and costly. Supplier relationship management varied: Alpha stays very close with proactive follow-up. Beta ensures visibility into use contexts to motivate performance, and Gamma monitors suppliers continuously for transparency and early interventions.

Customer influence on sourcing ranges from continuous to selective. In Alpha's case, the customer was deeply involved in projects and participated in sourcing decisions. Beta's customers participate at milestones and sometimes pre-define suppliers. Gamma keeps customers informed via integrated project teams but typically not in supplier choice, except where system safety is affected. Across cases, procurement mediates capability demands with cost, risk, and supply options. Organisational learning and requirement adaptation also differ. Alpha's mirrored structure aids technology transfer and iterative requirement refinement across long developments; Beta uses formal checkpoints to manage scope change; Gamma's monthly integrated teams surface risks and safety-critical component choices. Lifecycle continuity was evident in Alpha's processes (obsolescence management, advocacy for lifecycle obligations).

Prior research establishes internal integration as a precursor to external integration, enhancing knowledge sharing, process alignment, and production flexibility (Ataseven and Nair, 2017; Flynn et al., 2010). This study reveals a more complex bidirectional relationship in project-based environments. Our findings demonstrate that external customer and supplier integration can actively drive internal alignment, a dynamic that is less emphasised in existing literature. The Alpha case exemplifies how customer involvement shapes internal integration mechanisms and supply chain alignment, enabling procurement and production planning to respond to project-specific needs. Literature has primarily examined the role of procurement from an organisational structure perspective, emphasising formal roles, defined processes, and contractual mechanisms (cf. Pagell, 2004; Wynstra et al., 2003). This study, however, underscores the need to reconceptualise procurement as a dynamic actor within long-term, complex development projects. Rather than operating within predefined boundaries, procurement mediators must continuously navigate shifting project goals, evolving design choices, and emerging technology landscapes. This mediating role requires, for example, proactive engagement in cross-functional forums for technology road mapping (cf. Bhalla et al., 2022; Romero Rojo et al., 2010), adaptive coordination beyond initial project constraints, and strategic anticipation of supplier relationships (Hicks et al., 2000).

The analysis of procurement responses to external complexity, both dynamic and structural, indicates that it is important to consider adaptations and changes over time to support ETO-project completion, while also addressing cost concerns associated with structural factors. This raises several debates. Early sourcing of long-lead-time items, including customised parts, was crucial due to extended supplier development times. Consistent with Strandhagen et al. (2018), findings confirm that early ordering often occurs during the sales and preliminary specification process. However, this study adds nuance: such orders were often placed on a risk basis even before design was finalised, reflecting dynamic complexity. While procurement is typically associated with cost savings (cf. Murfield et al., 2021), the findings align with those of Leuschner et al. (2013), who observed no direct relationship between integration and cost savings, as higher integration often leads to increased coordination costs. This suggests that quality, lead-time, and specification fulfilment precede cost considerations (de Araújo et al., 2017; Kjersem and Giskeødegård, 2020), a dynamic complexity concern. Therefore, flexible, project-centric procurement structures are critical for ETO success, prioritising capability over cost (Gosling et al., 2020; Moretto et al., 2022).

ESI enhances ETO success by managing both dynamic and structural complexity through technical alignment, risk mitigation, and strategic sourcing. ESI provides early feedback and cost visibility, helping align expectations with feasibility (Cohee et al., 2018). Procurement-design collaboration ensures suppliers meet evolving cost and quality demands (Hicks et al., 2000; Reitsma et al., 2023). Dynamic complexity emerges in obsolescence management, requiring choices between last-time buys and switching suppliers (Romero Rojo et al., 2010). Single sourcing can reduce structural complexity but also limit the flexibility, especially under defence-related restrictions (Petersen et al., 2005).

Unlike Dixit et al. (2019), who emphasised early customer involvement, this study highlights the importance of continuous customer engagement. This approach enabled frequent trade-offs between lead-time and functionality, ensuring prioritisation aligned with evolving customer requirements. As shown in the findings, continuous customer integration contrasts with milestone-based approaches in other projects by allowing real-time adjustments to functionalities while managing shifts in lead-time. This highlights the importance of treating the customer as a co-creator of value (Dixit et al., 2019), particularly in high-dynamic complexity environments. This emphasises the case for expanding the traditional perceived boundaries of procurement beyond initial tendering and contracting, and for close and continual engagement with the customer.

Customer integration in Alpha extended beyond planning and into procurement and sub-projects. Customers provided knowledge that DefCo lacked, participated in supplier visits, and influenced sourcing decisions. This involvement helped manage complexity while presenting trade-offs between meeting customer preferences and leveraging industry expertise (Shurrab et al., 2022). The mirrored organisational structure for central engineering systems further facilitated knowledge transfer. This nuances the findings of Shurrab et al. (2022), showing that customer involvement in procurement can itself serve as a mechanism for managing complexity, albeit with new trade-offs.

Our research also highlights the challenges of setting accurate delivery dates in ETO environments due to long and variable lead-times, overlapping phases, product complexity, and uncertain capacity requirements, which is a central theme in ETO research (Cannas and Gosling, 2021; Gosling and Naim, 2009). The empirical findings show the dynamic and structural complexity factors that contribute to this, but also potential integration responses.

Table 4 outlines a comparison between the cases and reveals differences in complexity. Alpha faces the highest structural and dynamic complexity, characterized by strong early supplier and customer involvement, proactive obsolescence management, high organisational learning, and deeply embedded routines that handle complexity through integration and ownership. Beta shows medium complexity, relying on structured processes, staged supplier involvement, checkpoint-based management, and scope control to balance flexibility with control. Gamma, by contrast, experiences lower structural complexity due to limited customer influence but faces higher dynamic complexity, as it depends more on suppliers, milestone-based integration, and continuous adaptation, which introduces greater risks of change and reallocation delays.

The findings suggest that procurement in ETO environments requires not only effective integration but also an ability to tailor responses based on the type of complexity encountered. Overall, in structural complexity settings, the focus is on coordination, alignment, cost, supplier relationships, capabilities and life cycle management. Further, in dynamic complexity settings, the focus is on sourcing flexibility, adaptability, strategic partnerships and ongoing learning across supplier and customer interfaces. Table 5 summarises procurement activities in the integration dimensions and the two types of complexity – structural and dynamic.

Across internal integration and structural dynamics settings, procurement supports integration through procurement mediation roles embedded in projects, as well as cross-functional alignment mechanisms to manage technical interdependencies and ensure integration. Further, procurement supports integration through cost control and cost alignment early in the project, as well as acting as an internal interface mediator. Across internal integration and dynamic complexity settings, procurement enables responsiveness through cost-conscious design practices, early visibility of long lead-time items, obsolescence management through lifecycle guarantees and iterative coordination routines. External integration and structural dynamics rely on early supplier involvement, supplier capability alignment, and formal lifecycle continuity processes to handle long lead-times and complex supplier interfaces. Also, including customer mirroring the organization to support the flow of knowledge and enhance alignment. In external integration and dynamic complexity contexts, procurement adopts adaptive relationship management, requirement-driven system adaptation, and joint end-of-life planning to address evolving specifications. For example, Alpha, facing high structural and dynamic complexity with ∼1,000 suppliers and evolving requirements, embedded procurement mediators, on site customer participation, and formalised obsolescence management; Beta, with moderate complexity (≈120 suppliers), relied on a single procurement sourcing manager and milestone-based integration with design to cost controls; and Gamma, characterized by low structural but higher dynamic complexity, prioritised functionality and schedule over cost through stockpiling and contractual end-of-life protections.

In Figure 2, we position the studied projects within the structural and dynamic complexity framework. This framework distinguishes between four internal and four external integration logics based on inherent complexity. Each of these complexity configurations demand a tailored procurement response. The core characteristics of complexity and descriptors of procurement responses are included in the framework. Alpha indicates high structural and dynamic complexity, internally and externally, whereas Beta, exhibits moderate complexity on both dimensions, and Gamma lies toward the low–low end internally, while exhibiting intermediate complexity externally. Thus, procurement's integration role depends on the combination of structural and dynamic complexity, internally and externally. This aligns with prior conceptualisations that structural complexity is caused by the scope and interdependence of system elements, e.g. numerous components, large supply network, dense interfaces, whereas dynamic complexity is a result of uncertainty and frequent changes in requirements (Baccarini, 1996; Choi and Krause, 2006; Tatikonda and Rosenthal, 2000).

Across different internal-structural complexities, based on the number of interdependent parts and processes (Baccarini, 1996; Bozarth et al., 2009), the ETO-project reinforces stability by clarifying roles, instituting cross-functional coordination forums, and establishing cost-control and obsolescence management routines. To address internal structural complexity, Alpha embedded procurement into engineering early and clearly defined its role within project teams (Dowlatshahi, 1992; Reitsma et al., 2023). Beta used a lighter model with a single procurement coordinator mediating across functions, while Gamma relied on minimal internal integration, locking the scope early and reducing the need for heavy procurement involvement (cf. Moretto et al., 2022). Under internal-dynamic complexity, where requirements and designs are changing, procurement's role becomes more agile and adaptive (Azadegan et al., 2013; Maylor and Turner, 2017).

The emphasis is on rapid learning and flexibility: providing early visibility of long lead-time items (Guy and Dale, 1993; Petersen et al., 2005), safeguarding schedules through change-management (Iakymenko et al., 2022) and giving design teams leeway to adjust without losing control. Alpha illustrates this logic with frequent adjustments to evolving requirements, while Beta combined agility with cost management (design-to-cost) to maintain alignment under moderate uncertainty. Gamma, facing relatively fewer internal changes, only selectively engaged in adaptive practices, keeping internal processes lean. This dynamic internal integration logic reflects the need to manage uncertainty through responsiveness (Choi and Krause, 2006; Tatikonda and Rosenthal, 2000), ensuring that procurement can accommodate evolving project needs internally.

Under external–structural complexity, with a broad supplier base and technically complex system deliveries (Birkie and Trucco, 2016; Williams, 1999); procurement prioritizes formal and long-term integration measures. Key responses include securing early-supplier-involvement to lock-in critical components (Strandhagen et al., 2018), aligning supplier capabilities with technical requirements (Hicks et al., 2000; Reitsma et al., 2023), and implementing lifecycle continuity planning to mitigate risks like supplier lock-in or component obsolescence (Howard and Miemczyk, 2010; Romero Rojo et al., 2010). Alpha again represents the high–high case, anchoring supplier relationships early and proactively managing obsolescence. Beta applied similar measures but in a moderated form, balancing capability alignment and staged supplier involvement (Cannas and Gosling, 2021). Gamma, although smaller in scale, still had to secure continuity from a handful of single-source suppliers, illustrating how even low-structural-complexity projects require formal measures to avoid supply risk (Petersen et al., 2005).

On the other hand, external–dynamic complexity calls for more fluid integration. When external conditions are volatile, e.g. changing customer requirements or supply uncertainties, procurement leans on adaptive relationship management (cf. Laari et al., 2023; Picaud-Bello et al., 2022). In these contexts, integration is maintained through frequent supplier–customer communication, staged and intensive involvement at key milestones (Bhalla et al., 2022), and flexible contracting arrangements to accommodate shifting requirements. Alpha managed such volatility by embedding procurement closely with both suppliers and customers, enabling frequent interaction (Dixit et al., 2019). Beta used milestone-based reviews and stage-gated supplier involvement to balance structure and adaptability (Munir et al., 2020). Gamma, positioned closer to the intersection on the external side, relied heavily on adaptive relationships: frequent dialogue with suppliers, flexible contracts, and iterative adjustments to delivery schedules and specifications.

Crucially, these cases demonstrate that procurement must balance trade-offs and sequence integration responses as complexity unfolds (Maylor and Turner, 2017; Moretto et al., 2022). With finite time and resources, firms cannot maximally integrate on all fronts at once; procurement therefore assesses where integration efforts will have the greatest payoff given the project's dominant complexity at each phase (Baccarini, 1996). For example, Gamma's strategy of deferring intensive internal integration until after scope freeze and then ramping up external collaboration is a deliberate sequencing to match its complexity profile. Beta's concurrent but moderate internal/external integration reflects a trade-off to avoid over-commitment in either domain (Cannas and Gosling, 2021).

In practice, procurement should therefore play a continual mediating role sensing shifts in complexity and re-calibrating the integration approach accordingly (Laari et al., 2023). As requirements evolve or uncertainties are resolved, a project may transition between quadrants of the framework (e.g. from moderate toward higher complexity), prompting procurement to adjust its tactics and re-evaluate over time. This dynamic capability to reposition integration focus is pivotal in ETO environments (Tatikonda and Rosenthal, 2000), ensuring that procurement's efforts remain focused without overextending into unnecessary integration. Overall, the framework highlights procurement as a central mediator that enables hybrid integration configurations tailored to each project's complexity demands; reinforcing internal stability where needed, fostering external agility where possible, and balancing both to navigate the structural and dynamic challenges at hand.

The purpose was to examine procurement responses to varying complexities in ETO defence, asking: How can procurement respond to the integration of ETO supply chains across different types and levels of complexity? Complexity was conceptualised through definitions of dynamic, structural, internal, and external complexity, combined with a synthesis of SCI factors relevant to ETO defence. Procurement responses under different complexity combinations were proposed within a final framework to analyse their interplay.

This study provides insights into procurement's role in defence ETO-projects by integrating ETO complexity, SCI, and procurement literature. Bridging these areas advances understanding of procurement's mediating role in ETO defence. While prior research has acknowledged procurement's importance (Moretto et al., 2022), few studies have positioned it as a response to ETO-project complexity in achieving SCI (Cannas and Gosling, 2021). The interplay between SCI-dimensions, complexity characteristics, and procurement responses is shown in Figure 2.

We argue that complexity provides a basis for balancing integration efforts in procurement. Moving beyond the notion that increased integration is always better (cf. Frohlich and Westbrook, 2001; Ataseven and Nair, 2017), the findings suggest that procurement can be deliberately sequenced and prioritised for specific integration efforts depending on the balance and level of structural and dynamic complexity. Consequently, since managing complexity consumes significant costs, resources, time, and effort, it is essential to distinguish between these two forms. Structural complexity, such as the number of components and suppliers, remains a core theme in the SCI ETO literature, while dynamic complexity, which requires adaptation to technical and process changes over time (cf. Iakymenko et al., 2022), extends beyond traditional integration concepts. In this way, procurement is not only a facilitator but also a mediator of trade-offs, aligning scarce resources with the project's complexity profile.

Adding nuance to Dixit et al. (2019), we challenge the role of procurement as only an early-stage integrator, arguing instead for continued engagement throughout the project to mediate, facilitate, and integrate. We highlight that SCI must be continuously addressed not only upstream with suppliers but also downstream with customers, to achieve integration, an aspect less developed in prior ETO research (Cannas and Gosling, 2021). This contrasts with earlier, more limited views of procurement's role in ETO-projects (cf. Moretto et al., 2022), underscoring procurement's strategic influence on project outcomes. Similarly, while Ataseven and Nair (2017) argue that internal precedes external integration, our findings contrast with this linear view, showing instead that external customer and supplier engagement can trigger and reinforce internal integration. Finally, we extend contributions on obsolescence management (Howard and Miemczyk, 2010; Romero Rojo et al., 2010) by showing how procurement mediates lifecycle continuity as part of dynamic responses.

For managers in ETO environments such as project, design and procurement, the findings show that complexity can serve as a decision criterion for sequencing and balancing integration. Rather than pursuing maximal integration, managers should assess whether structural or dynamic complexities dominate in different phases and allocate resources accordingly. When structural complexity is high, formal role optimisation, supplier alignment, and lifecycle planning should be prioritised. Conversely, when dynamic complexity prevails, adaptive routines, flexible contracting and close customer–supplier dialogue are critical.

Based on the findings, we would encourage practitioners to firstly, ensure procurement and delivery teams remain closely linked to respond to changing integration needs over time. The tables and figures can be used at the project level as templates to guide collaborative, integrated planning meetings, as well as joint workshops. At a more aggregated level, they may be used proactively to help integrate teams in technology road mapping exercises with the inclusion of procurement professionals. Second, we emphasise the importance of understanding the mediating role of procurement within the overall management, and the critical role of customer engagement and involvement. Third, we propose a proactive, systematic and holistic view of complexity to analyse its impact and focus integration efforts. Tables 4 and 5 offer templates to consider procurement's mediating role, and Figure 2 offers a structure for reflecting on appropriate internal and external responses to various complexities.

This study is constrained by its single-case focus on defence-sector ETO projects within DefCo, a setting marked by high technical complexity, strict regulation and pronounced risk sensitivity (Gosling and Naim, 2009). While this context provides valuable insights into procurement integration under conditions of structural and dynamic complexity (Williams, 1999), its empirical generalisability to other industries remains underexplored. In sectors such as construction, shipbuilding, energy or IT, procurement may follow distinct logics, calling for cross-sectoral research to identify which integration mechanisms are context-specific and which hold more universal relevance. Future studies could therefore pursue comparative, cross-industry designs to validate and extend the framework, while longitudinal approaches would capture how procurement practices and their effects evolve across project life-cycles. Such research would not only strengthen external validity but also advance theoretical understanding of procurement's role in managing various levels of complexity within ETO supply chains.

The development of defence supply chains that can deliver at speed and scale is a societal undertaking. This study highlights the need for a research agenda focused on procurement's role in building robust defence supply chains. A central theme is how procurement can orchestrate SCI under conditions of structural and dynamic complexity (Maylor and Turner, 2017). Future studies should explore how procurement simultaneously aligns internal processes, multi-tier supplier networks, and international collaborations, particularly in cross-border initiatives such as those supported by the EU or NATO (Caldwell et al., 2009). Research is also needed on the inclusion of SMEs and of firms offering dual-use products, which remain underexplored despite their potential to extend capacity and innovation. Another important strand concerns resilience. Modern supply networks operate as adaptive systems where disruptions propagate rapidly (Choi et al., 2001), demanding procurement-led approaches that extend beyond conventional risk management. Further studies are required to investigate how procurement can balance speed, scalability, and sovereignty while enabling resilient integration in defence ETO contexts.