MOD-FCA: A Quantitative Reference Framework for Multi-Layered Closed-Loop Management Control in the Digital Era

Abstract

In the digital era, enterprises face increasing pressure to align strategic objectives with operational execution under volatile and data-intensive conditions. Traditional management control systems often rely on lagging indicators and ad hoc interventions, limiting both performance visibility and sustainability outcomes. This study develops MOD-FCA, a prescriptive, multi-layered closed-loop management control framework that links value-centric outcomes to business-centric drivers through vertically aligned metrics, objective tensors, and tiered corrective routines. Using a longitudinal case study in a manufacturing enterprise, we illustrate how MOD-FCA enhances operational traceability, supports systematic deviation identification and response, and institutionalizes organizational knowledge for continuous improvement. Importantly, MOD-FCA is designed to support sustainable industrial practices by embedding sustainability-related metrics, such as resource efficiency, energy intensity, waste reduction, and process compliance, into the same metric deployment, deviation triggering, and corrective-action logic used for operational control. Qualitative feedback from managerial and operational roles indicates that MOD-FCA strengthens accountability, ensures role-aligned responses, and fosters proactive, data-driven decision-making. These findings provide both theoretical contributions to management control system design and practical guidance for enterprises seeking to achieve both operational excellence and long-term sustainability.

1. Introduction

Amid the accelerating global digital economy, firms are increasingly exposed to VUCA conditions [1,2,3]. As demand cycles shorten and supply chains are disrupted more frequently, organizations must pair sound strategy with continuous execution monitoring, timely deviation correction, and dynamic alignment across multiple (often competing) objectives to improve goal-attainment certainty and enhance responsiveness [4,5,6]. Recent studies further suggest that digital transformation is reshaping performance measurement and management systems toward greater flexibility, real-time decision support, and sustainability orientation, while platform-based and networked manufacturing contexts intensify the need for cross-boundary coordination and governance [7,8,9].

These environmental pressures create three concrete management pain points that motivate this study. First, why do firms become data-rich but remain action-poor? Although operational data are increasingly available, they are often not translated into deviation signals, response routines, or closure evidence. Second, why are metrics and accountability frequently decoupled? Many firms monitor outcomes or completion rates, but the links among value outcomes, controllable process drivers, responsible roles, and corrective authority remain unclear. Third, why do deviations often lead to retrospective reporting rather than timely corrective action and learning? Without explicit thresholds, escalation rules, and knowledge codification, problems are commonly addressed through managerial experience and ad hoc interventions. These pain points suggest that digital-era Management Control System (MCS) design requires not only more data or more indicators, but also a prescriptive logic that connects metrics, objectives, deviations, responsibilities, and learning routines in a closed loop.

However, traditional management control often relies on periodic reporting and experience-based judgments, which are increasingly susceptible to cognitive biases in complex manufacturing environments [10]. Such practices commonly suffer from delayed feedback and inconsistent measurement definitions, leading to decision latency and information waste that are insufficient for the fine-grained and near-real-time governance required in the digital era [11,12]. Consequently, the question of how to redesign MCS to support data-driven operational governance and continuous improvement has re-emerged as a central issue in operations management research [1,13].

Prior studies suggest that digital technologies—such as the Internet of Things (IoT) and advanced analytics—are fundamentally reshaping key managerial processes, including planning, performance evaluation, and process monitoring [14,15]. In principle, these advancements can significantly enhance the transparency, timeliness, and predictive capability of MCS, moving them from reactive to proactive governance [16,17]. Consequently, the dual forces of increasing market complexity and technological enablers necessitate a systematic examination of how MCS architectures should evolve to match this dynamic competitive ecosystem [1,2].

MCS is essentially “the formal, information-based routines and procedures managers use to maintain or alter patterns in organizational activities” [18]. Its core function lies in translating strategic objectives into measurable indicators to guide operational alignment. However, when organizational digital maturity is insufficient, operational data acquisition remains limited and periodic (e.g., relying on manual monthly reports). This results in a dual deficiency of limited quantification and decision latency, where feedback loops are too slow to correct deviations in real time [4,19,20]. While foundational frameworks provide comprehensive templates of control elements [21,22,23], they function primarily as descriptive taxonomies rather than prescriptive frameworks. Consequently, enterprises still lack a generic, computable framework capable of supporting the vertical integration and real-time closed-loop governance required in the digital era [13,14].

To address this gap, this study adopts the Design Science Research (DSR) methodology [24,25] and builds on the MOD-FCA (Metric-Objective-Deployment-Feedback-Check-Action) logic recently introduced by Zhang, Bi [4] to develop a quantitative reference framework for multi-layered closed-loop management control in the digital era. Unlike descriptive taxonomies that primarily categorize control elements [22], MOD-FCA is positioned as a prescriptive reference artifact. It specifies the key constructs required for quantitative management control and provides a formalized design logic for (i) metric-system deployment, (ii) objective tensor setting, (iii) algorithmic deviation triggering, and (iv) knowledge codification for continuous improvement. The framework is intended to offer configurable guidance for organizations to build context-specific MCSs across different operational domains.

More specifically, the prescriptive nature of MOD-FCA lies in three mechanism-level differences from existing MCS/PMS frameworks. First, in terms of metric deployment, MOD-FCA does not stop at identifying performance dimensions; it specifies how value-centric outcomes are decomposed into business-centric driver metrics and then instantiated as objective profiles indexed by time and scenario. Second, in terms of deviation triggering, it converts plan–actual comparisons into computable deviation events through adaptive thresholds, allowing deviations to be classified and routed rather than merely reported. Third, in terms of accountability binding, MOD-FCA links each metric and deviation class to responsible roles, response routines, closure evidence, and knowledge codification. Thus, MOD-FCA complements descriptive MCS/PMS frameworks by providing an executable integration logic that connects metrics, objectives, deviations, responsibilities, and learning routines.

This study makes three specific contributions to the literature. First, regarding metric-system design, it proposes a method for designing and deploying a multi-layered metric system that emphasizes the structured decomposition of strategic objectives and consistency checking, ensuring vertical alignment between value outcomes and operational drivers [20]. Second, regarding the closed-loop governance mechanism, it introduces a tiered deviation-trigger-and-response mechanism based on adaptive thresholds. This embeds monitoring, diagnosis, and corrective action into formal routines, thereby shifting performance management from a passive “reporting system” to an active “action system” [26,27]. Third, regarding transferable evidence, the paper documents the instantiation process and observed outcomes of MOD-FCA through a longitudinal case study of Enterprise A. This offers traceable process evidence and constitutes a Level 2 knowledge contribution to the design of MCSs in data-intensive environments [28].

The remainder of the paper is organized as follows. Section 2 reviews and synthesizes research on MCS, performance management, and digitally enabled operational governance. Section 3 describes the DSR methodology and the research execution process. Section 4 presents the constructs, relationships, and operating mechanisms of MOD-FCA. Section 5 reports a longitudinal case study demonstrating instantiation and evaluation results. Section 6 concludes with theoretical implications and managerial insights.

2. Literature Review

2.1. The Evolution of MCS and PMS

The intellectual roots of MCS are commonly traced to Anthony’s [21] seminal work. Anthony conceptualized management control as the process by which managers ensure that resources are obtained and used effectively and efficiently to achieve organizational objectives, explicitly distinguishing it from strategic planning and operational control [21,29]. This hierarchical framing provided a clear lens for understanding control in bureaucratic organizations and, for decades, aligned closely with financially oriented measurement practices. Over time, however, researchers increasingly recognized that a strict separation between strategy formulation and operational control is difficult to sustain in dynamic environments—especially when control relies primarily on lagging financial outcomes rather than on the operational drivers that shape future performance [30,31].

In response, research in the 1990s expanded the boundaries of control. The Balanced Scorecard introduced non-financial measures and emphasized cause-and-effect linkages to translate strategy into operational terms. In parallel, Simons [18] proposed the Levers of Control framework, arguing that MCS should not be viewed solely as constraining mechanisms. In Simons’ view, MCS encompasses diagnostic control systems for monitoring critical performance variables and interactive control systems that stimulate learning and help organizations address strategic uncertainties [18,32]. Together, these contributions broadened the scope of MCS by explicitly foregrounding the tension between maintaining predictable performance and enabling strategic adaptation.

In the 2000s, the rise of the “performance management” perspective encouraged more integrative frameworks, reinforcing the overlap between MCS and performance measurement systems (PMS). Otley [33] argued for moving beyond a narrow focus on measurement toward a broader conception of performance management, proposing an analytical framework structured around objectives, strategies, targets, rewards, and information feedback. Building on this, Ferreira and Otley [22] expanded the approach into a twelve-question framework that incorporates mission, vision, and the “strength and coherence” among system elements. A central implication of this stream is that MCS/PMS are best examined as an interrelated set of elements—or a “package”—rather than as isolated tools [23,34].

Despite their value in identifying the key components of MCS (the “what”), these frameworks are primarily descriptive and diagnostic. Ferreira and Otley [22] explicitly position their framework as a heuristic for describing and analyzing systems rather than as a prescriptive design blueprint. Consequently, while the literature emphasizes strategic–operational alignment, it offers limited guidance on the “how”: how to specify traceable linkages across levels and translate strategic intent into computable control logic. This limitation aligns with recent calls to strengthen the design-oriented foundations of MCS research by moving beyond component lists toward formalized design logics capable of bridging theory and practice [35].

2.2. MCS as a Package or a System

A central debate in MCS research concerns whether control arrangements should be understood as a loosely assembled “package” or as an integrated “system.” Early studies often examined individual control practices—such as budgeting, planning, or performance evaluation—in isolation, but this reductionist approach has been criticized for overlooking the interactions among controls that coexist in organizations [36,37]. The package view, represented by Malmi and Brown [23], emphasizes that multiple controls—such as cultural, planning, cybernetic, reward, and administrative controls—often accumulate over time in response to local contingencies. This perspective is useful for describing what types of controls are present in organizations, but it also reveals a practical problem: accumulated controls may remain loosely coupled, generating inconsistent priorities, fragmented routines, and weak strategic–operational coherence.

The system view, by contrast, stresses complementarity and internal consistency among control elements. A control arrangement becomes a system only when the value of one control practice depends on and reinforces the design and use of others [34]. For instance, diagnostic metrics become more effective when they are complemented by clear accountability structures, corrective routines, and coherent incentives; otherwise, coexisting controls may function as independent components rather than mutually reinforcing mechanisms [38]. This distinction is particularly important in the digital era. As firms introduce dashboards, automated reports, IoT-based monitoring, and analytics tools, they may rapidly expand the “control package” without necessarily improving actionable decision-making. Without an explicit integration logic, more data and more control tools can still result in information overload, delayed responses, distorted reporting, or an illusion of control [22,39].

MOD-FCA responds to this package–system debate by providing an integration logic for the cybernetic and diagnostic cores of MCS. It does not claim to replace all cultural, administrative, or reward controls in the broader MCS package. Rather, it specifies how metrics, objectives, deviations, responsibilities, response routines, and knowledge accumulation can be connected into a multi-layered closed loop. In this sense, MOD-FCA helps transform fragmented control elements into a more coherent and actionable control system for digitally enabled operations. Its contribution is therefore not merely to add another control component, but to prescribe how key control elements can be configured and linked so that performance information can trigger timely response, closure, and learning.

2.3. Digital-Era Challenges: From Data Abundance to Actionability

Digital technologies such as the IoT, advanced analytics, and cloud platforms have dramatically expanded firms’ capacity to capture and store operational data. Yet, the promised benefits of digitalization do not automatically translate into effective management control. A recurring pattern reported in the literature is that organizations become data-rich while still struggling to generate actionable control information and timely managerial responses [14,40]. Despite having access to big data, many organizations face difficulties in translating raw data into meaningful insights due to a lack of holistic integration between social (people, goals) and technical (infrastructure, processes) elements [7,40]. The core problem is therefore not data scarcity, but the lack of a prescriptive logic that reliably connects operational data to managerial meaning and action.

First, data availability does not imply interpretability. Digital operations generate high-dimensional, heterogeneous, and noisy data. Without explicit semantic mapping and construct operationalization, raw sensor readings and transactional records remain detached from the managerial constructs on which control systems depend—such as deviations, risks, and performance gaps [17]. Furthermore, recent research highlights that “information flow waste”—such as unnecessary processing or quality defects—can impede the effective transformation of data into decision-making value [12]. When data are not reduced, structured, and prioritized before entering decision arenas, information overload can undermine rather than improve decision quality [41].

Second, conventional control practices often struggle to meet the real-time and cross-level requirements of digital operations. Earlier studies noted that performance measurement systems dominated by lagging, periodic reporting tend to provide “rear-view” control, limiting timely corrective action [19]. More recently, even when advanced digital systems are deployed, weak interoperability and fragmented data architectures can create persistent gaps between operational events and managerial visibility. You and Wu [42] note that in many construction and manufacturing contexts, ERP and BIM systems remain isolated, leading to manual data transmission and limited traceability. Under such conditions, digital systems risk becoming post hoc recording tools rather than enablers of closed-loop control.

Third, data quality and contextualization remain critical bottlenecks. Inaccurate, incomplete, or poorly maintained input data have been identified as a primary barrier to effective (and intelligent) control, often caused by the absence of standardized work procedures or operator negligence [43]. More fundamentally, when data are not explicitly linked to control context—such as targets, thresholds, responsibilities, escalation rules, and decision routines—they cannot reliably trigger appropriate corrective action. This is consistent with digital maturity perspectives, suggesting that connectivity alone is insufficient; structured sharing, interpretation, and governance are necessary to translate data into control capability [44].

In sum, the central challenge of management control in the digital era is not to collect more data, but to establish a prescriptive transformation logic that connects heterogeneous, high-frequency operational data to multi-layered targets, deviation evaluation, and corrective-action mechanisms. This motivates the need for a quantitative reference framework—such as MOD-FCA—that specifies how data streams can be translated into actionable control logic and embedded into closed-loop governance routines [1].

2.4. Cybernetic Logic and the Upgrade of Diagnostic Control

A useful lens for addressing the data–action disconnect is the cybernetic view of organizational control. Otley and Berry [45] conceptualize control as an information feedback process and identify four necessary elements: (i) predefined objectives, (ii) measurement of actual performance, (iii) a comparator that evaluates deviations from targets, and (iv) an effector that enables corrective action. From this perspective, effective management control is not merely about information reporting; it is a closed-loop arrangement in which deviations are detected, interpreted, and addressed in a timely and disciplined manner.

Within mainstream MCS frameworks, diagnostic control systems represent a prominent managerial instantiation of this cybernetic logic. Diagnostic control relies on exception-based monitoring to compare outcomes against preset standards and to intervene when significant deviations occur, thereby economizing managerial attention [18,32]. However, research has long noted that diagnostic systems can become overly static in dynamic environments—particularly when performance standards, operational constraints, and sources of variation change faster than the review and adjustment cycle [19]. In such contexts, diagnostic control risks degenerating into lagging, report-oriented monitoring rather than enabling timely corrective action.

Digital technologies create opportunities to strengthen closed-loop control, but they also raise the bar for control design. Ferreira and Otley [22] emphasize that contemporary performance management should incorporate not only feedback information based on observed deviations (enabling single-loop learning) but also feed-forward information that supports anticipation and strategic questioning (enabling double-loop learning). While feedback loops correct actions to meet given targets, feed-forward mechanisms support questioning the validity of the underlying strategy and the adaptation of targets and thresholds [22,26].

Despite these opportunities, the literature suggests that many organizations’ digitally enabled control practices remain only weakly “closed-loop.” Rear-view control persists, and translating high-frequency data into timely, structured responses remains difficult [20]. To address this, recent scholarship advocates viewing performance management through a “System of Systems” (SoS) perspective, in which autonomous control loops at different levels must be interconnected to ensure alignment amid complexity [27]. Effective control requires adaptive capability: beyond reporting deviations, the control arrangement must trigger appropriately tiered responses, clarify responsibilities, and embed diagnosis into formal routines [27,46]. This motivates the development of MOD-FCA as a prescriptive reference framework that operationalizes cybernetic logic into a multi-layered, quantitative closed-loop system suitable for the digital era.

2.5. Research Gap

The literature reviewed above has substantially advanced our understanding of MCS in terms of core elements, systemic configurations, and digital-era challenges. Yet, a critical meso-level design gap remains between high-level conceptual frameworks and the specific requirements of digitally enabled operations. Recent scholarship notes that the academic literature often lags behind leading-edge practice, failing to provide the explanatory or prescriptive logic necessary to bridge the widening gap between theoretical constructs and the operational reality of digital manufacturing [35,47].

First, foundational MCS frameworks are predominantly descriptive, diagnostic, or taxonomic. Ferreira and Otley [22], for instance, explicitly position their framework as a “heuristic tool” to facilitate the description and analysis of system design and use, rather than as a normative blueprint for system design. Similarly, while Malmi and Brown’s [23] typology provides a structured classification of control practices (e.g., cultural, planning, cybernetic), it functions primarily as a checklist of what exists, offering limited guidance on how to achieve coherence across these elements in a specific context. As a result, practitioners often possess a vocabulary of control components but lack the design rules required to specify traceable cross-level linkages and translate strategic intent into deployable digital logic.

Second, while systemic perspectives emphasize interdependence and internal consistency as defining attributes of effective control systems [34], the literature provides comparatively limited guidance on how such coherence can be engineered. Specifically, there is a lack of formalized methods for linking quantitative targets, responsibilities, and response routines so that deviations reliably trigger appropriate managerial action. This limitation is particularly salient in data-intensive environments, where the abundance of operational data heightens the need for an explicit logic to translate raw signals into structured control decisions, preventing information overload from paralyzing the system [14].

Third, the shift from static monitoring to dynamic closed-loop control calls for design principles that embed both feedback and feed-forward information into formal routines. However, existing contributions remain fragmented across conceptual discussions and domain-specific tools, leaving a gap in prescriptive knowledge that integrates management logic with digital information flows [1,20].

Table 1. Positioning of MOD-FCA relative to representative MCS/PMS and digital control frameworks.

Framework	Main Focus	Limitations for Digital Closed-Loop Control	MOD-FCA’s Incremental Contribution
Anthony’s management control hierarchy	Distinguishes strategic planning, management control, and operational control.	Provides a high-level hierarchy but limited guidance on computable cross-level deployment and real-time deviation response.	Extends hierarchical control into four operationalized layers with explicit metric deployment, comparator logic, and tiered responses.
Balanced Scorecard	Translates strategy into financial and non-financial measures.	Emphasizes strategic alignment but does not prescribe how deviations trigger routines or how accountability is bound to response paths.	Links value-centric metrics to business-centric drivers and embeds them into objective tensors and response routines.
Levers of Control/diagnostic control	Uses critical performance variables for exception-based monitoring.	Provides a conceptual logic for diagnostic monitoring but is less specific about digital event classification and closure mechanisms.	Formalizes diagnostic control through adaptive thresholds, event classes, and self-correction/escalation rules.
MCS package/system perspective	Emphasizes complementarity and coherence among control elements.	Explains the importance of coherence but provides limited construction rules for engineering such coherence.	Provides an integration logic that binds metrics, objectives, responsibilities, routines, and knowledge accumulation.
Digitally enabled PMM/analytics-based control	Uses digital data, analytics, dashboards, or predictions to improve visibility.	Often improves data visibility but may remain weak in translating signals into organizational action.	Converts data into deviation events and connects them to role-aligned corrective routines and feed-forward updates.

presents the positioning of MOD-FCA. Therefore, the contribution of MOD-FCA is not to replace existing MCS/PMS theories but to operationalize their core insights into a mechanism-oriented reference framework suitable for data-intensive and digitally enabled operations.

3. Research Approach

We selected DSR as the research method, leveraging its strengths in designing artifacts to solve practical problems in fields such as Information Systems [24] and Decision Science [48]. This study aims to design a reference framework to support enterprises in designing refined management control systems. DSR focuses on knowledge-intensive design and addresses practical problems through the development of innovative artifacts, with demonstrated effectiveness in the management field [24]. Furthermore, DSR complements behavioral science: DSR expands human and organizational capabilities by creating novel artifacts, while behavioral science develops and validates theories that explain and predict human and organizational behavior [48]. We employ the Design Science Research Process (DSRP) model as our guiding framework. The model is consistent with established DSR protocols. In accordance with Peffers, Tuunanen [25], Figure 1 illustrates the six steps of DSRP.

Step 1 (Problem Identification and Motivation) requires defining the research problem, clarifying its scope, and justifying the value of a solution. Since the problem definition will be used to develop an artifact that provides an effective solution, conceptually decomposing the problem may help ensure that the resulting artifact adequately addresses the complexity of the problem [25]. This step helps engage stakeholder interest, allowing them to understand the researcher’s problem-solving logic, while also aiding in understanding the reasoning underlying the research problem. This study identifies that advanced and effective management control systems are key for enterprises to develop competitive advantages and cope with intense market competition [13]. However, in the digital era, how should MCSs evolve to adapt to the development of information technology? Enterprises still lack an effective framework reference in practice, making it difficult to establish scientific and sustainable control systems [4].

Step 2 (Definition of Solution Objectives) requires defining the objectives for a solution based on the inferences from the previous step. The objectives should be rationally derived based on the problem statement. Required resources include knowledge of the current state of the problem, existing solutions (if any), and their effectiveness [25]. This study aims to enhance the sophistication of enterprise management control by developing a generic reference framework that helps enterprises establish refined management control systems. Based on an in-depth analysis of enterprise management needs in the digital era, we further specify the objectives into the following three design goals: (D1) Develop a method for designing a vertically aligned quantitative indicator system; (D2) Design a closed-loop control mechanism with real-time sensing, diagnosis, and dynamic adaptation; (D3) Ensure the framework’s genericity and extensibility. This objective system essentially represents a systematic adaptation and strategic upgrade of the original control system framework in the context of the digital era.

Step 3 (Design and Development) involves designing and developing the artifact that solves the research problem. Such artifacts can be constructs, models, methods, or instantiations (all broadly defined), or “new properties of technical, social, and/or informational resources” [24]. This stage is the core construction process of the MOD-FCA framework. We integrated Anthony’s three-level model, the feedback principle from control theory, and the PDCA cycle to form the conceptual prototype of the framework. Subsequently, we introduced formal modeling methods and designed a formalized framework specification. This stage involved three iterative versions. For example, in version V2, based on expert feedback, we refined the initial “three-layer” hierarchy into a “four-layer” one to more clearly separate management and control functions, enhancing the framework’s practical applicability and organizational adaptability. This refinement was not intended to add organizational hierarchy, but to make the comparator function explicit by separating managerial translation from deviation comparison and response activation.

In Step 4 (Demonstration), the researcher must demonstrate the use of the artifact to solve one or more instances of the problem. Common demonstration methods include experiments, simulations, case studies, and prototype testing. This study employs a mixed-methods approach combining quantitative and qualitative research, selecting a casting enterprise as the case organization for an in-depth longitudinal case study [49]. The case not only provided a testing ground for the MOD-FCA framework in a real business environment, but the research process itself also continuously fed back into the framework’s refinement, driving its ongoing optimization. Notably, the demonstration phase of MOD-FCA overlapped significantly and proceeded concurrently with the design and development phase (Step 3). Although the methodology requires maintaining consistency with existing project data during the research process, real-time feedback from frontline participants allowed us to continuously calibrate and enhance the framework within the dual cycles of development and validation. This process thus evolved into an iterative exploration, which not only helped find an effective solution to the core problem driving this research but also profoundly embodies the fundamental purpose of DSR—to create artifacts whose utility is demonstrated to both the academic and practitioner communities [24].

In Step 5 (Evaluation), the aim is to systematically observe and measure the actual effectiveness of the artifact in solving the target problem, focusing on comparing intended objectives with observed outcomes during the demonstration phase. Depending on the specific problem context and artifact characteristics, evaluation can take various forms, including: comparing artifact functionality against design objectives, analyzing objective quantitative indicators such as key budgets and outputs, user satisfaction surveys, in-depth customer feedback, testing in simulated environments, or performance indicator evaluations such as system response time and usability [25]. This study employs both formative and summative evaluation. Among these, formative evaluation runs throughout the entire demonstration phase, focusing on the refinement and improvement of the framework; whereas the summative evaluation centers on the in-depth casting enterprise case described in Part 5, systematically verifying whether the MOD-FCA framework fully achieves its various design goals set in Step 2.

In Step 6 (Communication), the researcher needs to disseminate the research results to the academic community and other relevant audiences (e.g., industry practitioners). To date, we have presented and discussed MOD-FCA with both industry and academia (e.g., at the China FAW Group Second “Qizhi·iM” Advanced Manufacturing Technology High-End Forum, and the Jilin Province Digital-Intelligent Transformation 20-Person Forum).

4. Artifact Description (Proposed MOD-FCA Framework)

Section 4.1, Section 4.2, Section 4.3 and Section 4.4 present the generic artifact design of MOD-FCA independently of the case context. The enterprise case is reported separately in Section 5 as an instantiation and evaluation of this generic design. The artifact combines (i) a structural model that specifies layers, interfaces, and information flows, and (ii) an operational method that governs how metrics, objectives, and corrective actions are deployed and executed.

4.1. Overview

MOD-FCA is grounded in a cybernetic view of control, in which objectives, measurement, comparison, and correction constitute a closed-loop arrangement across organizational levels [45]. Figure 2 summarizes the framework as comprising (i) four organizational layers that differ in information granularity and control cadence, extending Anthony’s [21] classic hierarchy; (ii) two coupled logic flows—Metric-Objective-Deployment (MOD) and Feedback-Check-Action (FCA)—that connect these layers bidirectionally, to ensure vertical alignment [50]; and (iii) a supporting digital foundation that enables traceability and routine execution [12]. Formal symbols and interfaces referenced in Figure 2 are defined in Section 4.2, while the operational procedures for deploying and operating the framework are detailed in Section 4.3.

4.1.1. Layering and Control Entities

To support vertical integration between strategic intent and operational execution, MOD-FCA differentiates control activities into four layers based on information granularity and control cadence. The rationale for layering is practical and theoretical: in many organizations, managers either disengage from operational governance or become absorbed in frontline firefighting, and the organization struggles to ensure that resources are obtained and used effectively and efficiently to achieve objectives. A multi-layered architecture clarifies division of labor and responsibility boundaries, enabling tiered cadences and role-aligned accountability consistent with cybernetic control logic [45].

The four-layer configuration should be understood as a functionally sufficient and parsimonious architecture rather than as a universally optimal number of organizational levels. Its rationale lies in separating four core control functions that are often conflated in traditional hierarchical control: value orientation, resource and governance translation, deviation comparison and response activation, and operational execution/data generation. A three-layer structure, such as the classic strategic–managerial–operational hierarchy, is useful for describing authority levels, but it tends to merge the comparator function into either management or operations. In practice, this may lead to delayed deviation recognition, inconsistent escalation, or self-reporting bias, because the same actors who set targets or execute tasks are also expected to judge deviations and trigger corrective routines.

By introducing a distinct control layer, MOD-FCA separates “doing” from “checking and triggering.” The control layer consolidates measurements, evaluates deviations against deployed objectives, classifies event severity, and activates predefined response routines. This makes the cybernetic comparator explicit and auditable. At the same time, a five-layer or more fine-grained structure would add organizational complexity without necessarily adding a new essential control function at the reference-framework level. Therefore, the four-layer structure balances theoretical completeness and practical parsimony: it preserves strategic direction, managerial coordination, independent control orchestration, and operational execution as distinct but connected functions.

Strategic layer. Defines value orientation and strategic objectives, specifying the high-level value-centric metrics used to steer the organization. Decisions at this layer are typically low-frequency (e.g., quarterly or annual) and focused on long-term adaptation.

Managerial layer. Translates strategic objectives into business-unit priorities, resource allocations, and governance routines. It provides the organizational bridge between strategic direction and operational control, ensuring that local optimization aligns with global goals.

Control layer. Functions as a dedicated comparator and control-orchestration role distinct from both managerial target/resource allocation and operational execution. It consolidates measurements, evaluates deviations against deployed targets, classifies event severity, and triggers exception-handling routines according to predefined rules. This layer ensures consistency and timeliness in deviation identification and response activation, while reducing the risks of self-reporting bias, delayed escalation, and ad hoc firefighting.

Operational layer. Executes standard operating procedures (SOPs) and generates primary process data (e.g., sensor readings, system logs, and structured records). This forms the empirical basis for feedback, leveraging Industry 4.0 technologies for real-time visibility [51].

4.1.2. Coupled Logic Flows

MOD-FCA links layers through two complementary flows. Their design responds to three recurring control challenges: (i) indicator systems dominated by completion rates or outcomes that fail to measure capability and therefore cannot reliably reveal or trace problems; (ii) heavy dependence on inspections that are costly and rarely achieve full coverage, alongside incentives that can suppress truthful problem exposure; and (iii) weak closure discipline in which even visible problems are handled through supervisory pressure rather than routine-driven corrective action.

• MOD (downward). MOD specifies how value-centric objectives are translated into business-centric metrics and deployable targets at lower layers. The output of MOD is a structured target set and associated measurement definitions that enable consistent monitoring across layers. The rationale is to make control objects explicitly quantitative: by linking value outcomes to controllable drivers, MOD improves problem visibility and traceability, transforming outcome indicators into diagnostic and actionable measures.
• FCA (upward). FCA specifies how operational data are aggregated into performance observations, compared against deployed targets, and transformed into decision-relevant deviation signals that trigger appropriate governance routines. FCA therefore strengthens proactive feedback by treating deviations as explicit control objects and by embedding reporting and monitoring into formal routines—reducing dependence on exhaustive inspection and shortening problem discovery cycles. Importantly, FCA is not limited to reporting deviations; it specifies the logic for closed-loop correction by linking deviation signals to tiered responses (self-correction vs. escalation) and to closure requirements, thereby shifting performance management from a reporting system toward an action system.

The two flows are mutually dependent: MOD determines what should be monitored and achieved, whereas FCA determines what is occurring, whether deviations exist, and which routines should be activated for correction and adaptation.

4.1.3. Supporting Elements

MOD-FCA assumes a digital foundation that enables data integration, traceability, and the execution of formal routines. In particular, the knowledge repository serves as organizational memory by maintaining metric definitions, SOP references, rule parameters (e.g., thresholds), and historical deviation–response cases. The rationale for knowledge codification is that operational know-how often remains tacit and person-dependent; without institutional memory, organizations repeatedly encounter similar problems and rely on experienced individuals rather than stabilized routines. By accumulating reusable deviation–response patterns and parameter histories, the repository supports consistent deployment, reduces repeated firefighting, and enables continuous refinement of targets, thresholds, and routines over time.

4.2. Formal Definition of Constructs

To reduce ambiguity in conventional PDCA descriptions and enable computational deployment, MOD-FCA specifies a compact set of constructs and operators. The framework can be represented as interactions among (i) metric vectors, (ii) multi-dimensional objective (target) structures, (iii) deviation measures, and (iv) adaptive thresholds that govern escalation and correction.

Notation (core symbols). Let $l\in \mathcal{L}$ denote the organizational layer (e.g., strategic/managerial/control/operational), $i\in \{1,\dots ,n_l\}$ the metric index at layer $l$, $t\in \{1,\dots ,T\}$ the time index (rolling control cycles), and $s\in \{1,\dots ,S\}$ the scenario index (e.g., baseline vs. challenge).

⮚

$M^{\left(l\right)}\in {\mathbb{R}}^{n_l}$: metric vector at layer $l$.

⮚

$O^{\left(l\right)}\in {\mathbb{R}}^{T\times n_l\times S}$: objective tensor (time × metric × scenario).

⮚

$x_t\in {\mathbb{R}}^{n_l}$: performance vector observed at time $t$.

⮚

${\delta }_t\in {\mathbb{R}}^{n_l}$: elementwise deviation rate at time $t$.

⮚

${\Delta }_t\in {\mathbb{R}}_{\ge 0}$: aggregate deviation magnitude.

⮚

${\mathsf{\Phi }}_j(t)$: adaptive threshold for metric $j$ at time $t$.

4.2.1. Metric Vector Space

At each organizational layer $l$, the metric system is defined as a vector:

$$M^{\left(l\right)}=(m_1^{\left(l\right)},\hspace{0.25em}{m}_2^{\left(l\right)},\hspace{0.25em}\dots ,\hspace{0.25em}{m}_{n_l}^{\left(l\right)}{)}^{\mathrm{\top }}$$

(1)

where each component $m_j^{\left(l\right)}$ denotes a quantifiable indicator (e.g., first-pass yield, SOP compliance rate). Vertical alignment is represented by a deployment mapping that links metrics across adjacent layers:

$$M^{(l;1)}=f_{\mathrm{deploy}}\left(M^{\left(l\right)}\right|\hspace{0.25em}{W}^{\left(l\right)})$$

(2)

where $W^{\left(l\right)}$ denotes an influence/weight structure capturing how higher-level priorities are translated into lower-level metric definitions and emphases. The detailed design procedure for $f_{\mathrm{deploy}}$ is specified in Section 4.3.

4.2.2. Objective Tensor

To avoid static target setting and to support rolling control under varying operating conditions, MOD-FCA uses a third-order objective tensor:

$$O^{\left(l\right)}=\left\{O_{t,j,s}^{\left(l\right)}\right\}$$

(3)

Here $O_{t,j,s}^{\left(l\right)}$ denotes the target value of metric $j$ at time slice $t$ under scenario $s$. The three dimensions represent:

Time ($t$): enabling rolling targets rather than fixed annual goals.

Metric ($j$): aligned with the metric index in $M^{\left(l\right)}$.

Scenario ($s$): alternative control baselines (e.g., baseline vs. challenge) that can be switched without redesigning the entire system.

4.2.3. Performance Vector and Feedback Mapping

The feedback module maps raw operational data into a structured performance vector that is comparable with the deployed objectives:

$$x_t=(x_{t,1},\hspace{0.25em}{x}_{t,2},\hspace{0.25em}\dots ,\hspace{0.25em}{x}_{t,n}{)}^{\mathrm{\top }},x_t=g_F(d_t)$$

(4)

where $d_t$ denotes the raw data at time $t$(e.g., sensor signals, system logs, or structured manual inputs), and $g_F\left(·\right)$ denotes the mapping and preprocessing logic that produces metric-consistent observations. This makes “feedback” a computable input to control, rather than informal communication.

4.2.4. Deviation Logic and Adaptive Thresholds

For a given scenario $s$, the instantaneous deviation rate of metric $j$ at time $t$ is defined as:

$${\delta }_j\left(t\right)={\displaystyle \frac{x_{t,j}-o_{t,j,s}}{o_{t,j,s}}}$$

(5)

To summarize multi-metric deviations, an aggregate deviation magnitude is defined as a weighted norm:

$${\Delta }_t=\sqrt{{\displaystyle {\sum }_{j=1}^N}{w}_j\left(t\right)\times {{[\delta }_j\left(t\right)]}^2},{\displaystyle {\sum }_{j=1}^n}{w}_j=1,\hspace{0.25em}{w}_j\ge 0$$

(6)

Rather than using fixed tolerance rules, MOD-FCA adopts metric-level adaptive thresholds ${\mathsf{\Phi }}_j(t)$ based on recent volatility and control criticality:

$${\mathsf{\Phi }}_j\left(t\right)={\mathsf{\Phi }}_0\times \left(1+{\alpha }_1\times \left({\displaystyle \frac{{\sigma }_{history}}{{\sigma }_{base}}}\right)\right)\times \left(1+{\alpha }_2\times C_{stratgy}\right)$$

(7)

where ${\mathsf{\Phi }}_0$ is a base threshold (initial tolerance), ${\mathsf{\sigma }}_{\mathrm{history}}$ is historical volatility (e.g., rolling standard deviation), ${\mathsf{\sigma }}_{\mathrm{base}}$ is a normalization baseline, ${\mathrm{C}}_{\mathrm{strategy}}\in \left[0,1\right]$ captures strategic criticality, and ${\mathsf{\alpha }}_1,{\mathsf{\alpha }}_2$ are sensitivity coefficients.

For system-level triggering, two dynamic bands can be set as:

$${\mathsf{\Phi }}_{\mathrm{low}}\left(t\right)=\mathsf{\Phi }\left(t\right),{\mathsf{\Phi }}_{\mathrm{high}}\left(t\right)=\gamma \hspace{0.17em}\mathsf{\Phi }\left(t\right),\hspace{0.25em}\gamma >1$$

(8)

These bands implement the “within/exceeds threshold” gate shown in Figure 2 and provide the basis for classifying event severity.

4.2.5. Hierarchical Response Mechanism

Given ${\Delta }_t$ and the dynamic bands, the control logic outputs an event class and triggers a tiered response function:

$$R\left({\Delta }_t\right)=\{\begin{array}{cc}\mathrm{Null},& \mid {\Delta }_t\mid <{\mathsf{\Phi }}_{\mathrm{low}}\left(t\right)\\ \text{Self-Correction},& {\mathsf{\Phi }}_{\mathrm{low}}\left(t\right)\le \mid {\Delta }_t\mid <{\mathsf{\Phi }}_{\mathrm{high}}\left(t\right)\\ \mathrm{Escalation},& \mid {\Delta }_t\mid \ge {\mathsf{\Phi }}_{\mathrm{high}}\left(t\right)\end{array}$$

(9)

This rule set operationalizes the middle block in Figure 2 (“plan–actual comparison & event classification”) and directly maps to the tiered responses in FCA:

Operational response: local corrective action for contained deviations (often triggered under Self-Correction depending on context).

Control response: root-cause check + action plan; closure records (Self-Correction).

Managerial response: cross-unit coordination/resource reallocation (Escalation).

Strategic response: target/threshold adjustment and policy update (Escalation + feed-forward update).

Accordingly, the FCA module produces three output types consistent with Figure 2: (i) corrective actions and closure records, (ii) updated thresholds/targets (feed-forward), and (iii) improvement cases stored in the knowledge repository.

4.2.6. Illustrative Calculation (Single-Metric Example)

Consider metric $j=3$ representing SOP execution compliance. For scenario $s$, suppose:

$$O_{t,3,s}=0.98,x_{t,3}=0.92$$

(10)

Then the deviation rate is:

$${\delta }_3(t)={\displaystyle \frac{0.92-0.98}{0.98}}\approx -0.061$$

(11)

Given the current threshold bands ${\mathsf{\Phi }}_{\mathrm{low}}\left(t\right)$ and ${\mathsf{\Phi }}_{\mathrm{high}}\left(t\right)$, the comparator classifies the event according to Equation (9) and triggers the corresponding tiered routine (self-correction or escalation). The generated event record $e_t$ includes the deviation magnitude/class and contextual descriptors (e.g., line, product family, shift, responsible role), enabling traceable closure and subsequent learning in the knowledge repository.

4.3. Operational Procedure

4.3.1. The MOD Module

The MOD module specifies how strategic intent is translated into a deployable, multi-layered metric and target structure. Rather than a simple KPI cascade, MOD is a prescriptive design-and-deployment procedure that produces three standardized artifacts (Figure 3): Output 1 a metrics dictionary and linkage structure (V-metrics and B-metrics), output 2 time- and scenario-indexed target profiles instantiated as an objective tensor $\mathrm{O}\in {\mathbb{R}}^{T\times N\times S}$, and output 3 a responsibility–routine map (who monitors/who responds).

(A) Value-centric metrics design (strategic layer)

Value-centric metrics (V-metrics) capture stakeholder value realization and represent the outcomes that the organization ultimately seeks (e.g., on-time delivery rate, customer complaints). They address the question: “Have we achieved the intended strategic outcomes?” MOD designs V-metrics through a traceability chain linking stakeholders, needs, and measurable outcomes:

A1 Stakeholder mapping. Identify key stakeholder groups and their relative salience using established logic (e.g., power–interest positioning).

A2 Needs elicitation. Elicit and consolidate stakeholders’ needs using mixed evidence (interviews, internal documents, surveys), producing a structured inventory of needs.

A3 Metrics conversion (candidate V-metrics). Translate prioritized needs into measurable outcome indicators and define calculation rules to ensure measurability, interpretability, and stable measurement definitions.

A4 Needs prioritization. Rank needs to be used with a structured weighting approach (e.g., AHP or weighted scoring). Weights are treated as context-calibrated parameters and are documented for replicability.

A5 Final V-metrics (strategic layer). Consolidate the top-ranked needs into a compact V-metric set. Each selected V-metric must be (i) outcome-oriented, (ii) measurable with explicit data sources, and (iii) actionable in the sense that it can be influenced by controllable process drivers at lower levels.

(B) Business-centric metrics structuring (managerial/control/operational layers)

Business-centric metrics (B-metrics) represent the process drivers of value creation at an appropriate level of operational granularity (e.g., schedule adherence, first-pass yield, SOP compliance, changeover time). They answer: “How is value created and controlled within processes?” MOD structures B-metrics using business process analysis (BPA) and a coverage discipline:

B1 Deliverables identification. Identify measurable deliverables that materially determine value outcomes (products, service outputs, compliance artifacts, operational plans). Deliverables must be defined in quantifiable terms so they can be monitored.

B2 Six-dimensional coverage. Select B-metrics to ensure balanced coverage across the following six facets (VQCECD, Volume-Quality-Cost-Efficiency-Compliance-Digital Maturity), preventing one-dimensional optimization. The purpose is not to maximize metric count, but to ensure that the major performance facets relevant to the value objective are not omitted.

B3 Candidate B-metrics (process/sub-process/activity). Decompose operations into process → sub-process → activity levels and define corresponding metrics at each level. This anchors measurement in controllable routines rather than detached reporting.

(C) Linkage & verification gate (BM → VM)

After A and B, MOD establishes explicit driver–outcome linkages so that alignment becomes operational rather than rhetorical.

C1 Build causal linkage candidates. Propose candidate linkages from B-metrics (drivers) to V-metrics (outcomes) based on process logic and managerial knowledge (e.g., schedule adherence and rework rate as drivers of on-time delivery).

C2 Verification gate (keep/drop links). Validate and refine linkages using one or more of the following mechanisms; only validated linkages are retained:

1:1 inheritance: the upper-level metric is directly computable by aggregation from a lower-level metric with a consistent definition.

1:N decomposition: the upper-level metric is structurally decomposed into multiple lower-level drivers (e.g., delivery performance decomposed into schedule adherence, rework rate, and dispatch accuracy).

1:M predictive linkage: historical data and/or structured expert review support the contribution of multiple lower-level metrics to an upper-level outcome (e.g., correlation/regression analysis and logic-based validation).

MOD outputs and handoff to FCA (Input A: planned)

Based on A–C, MOD produces three standardized outputs (Figure 3):

Output 1: Metrics dictionary & linkage structure (V-metrics and B-metrics). Includes the V- and B-metric dictionaries plus the validated linkage structure used for consistent cross-level deployment and interpretation.

Output 2: Target profiles instantiated as the objective tensor $\mathrm{O}(T\times N\times S)$. MOD generates time- and scenario-indexed target profiles and instantiates them as the objective tensor $\mathrm{O}$ (Equation (3)). The procedure includes:

• Time indexing: translating strategic targets into rolling windows aligned with governance cadence (e.g., monthly review with weekly monitoring).
• Scenario indexing: specifying baseline/stretch (and, when needed, guardrail) scenarios to enable controlled switching without redesign.
• Cascading & consistency constraints: deploying targets from higher to lower levels using the validated linkage structure so local targets remain consistent with upper-level intent.

Output 3: Responsibility–routine map (who monitors and who responds). For each metric (and for each event class where needed), MOD assigns monitoring responsibility, response responsibility, and the routine binding (which formal routine is triggered under self-correction vs. escalation).

Together, Outputs 1–3 constitute the planned interface consumed by the comparator in Section 4.2, enabling deviation detection, event classification, and tiered response triggering in the FCA loop.

4.3.2. The FCA Module

The FCA module operationalizes the bottom–up closed-loop logic in Figure 2. It converts operational data into a structured deviation event and triggers tiered responses across organizational layers. FCA is executed in recurring control cycles, and each cycle produces standardized records that support traceability, closure, and learning.

Step F1: Feedback—construct the performance vector ${\mathrm{x}}_t$

At the control cycle $t$, raw operational data ${\mathrm{d}}_t$ (sensor streams, system logs, inspections, and structured manual inputs) are mapped into the metric-consistent performance vector ${\mathrm{x}}_t$ using the feedback mapping defined in Equation (4). This step ensures that the “actual” input is semantically aligned with the deployed metric dictionary (Output 1 in MOD) and comparable with the planned objective tensor $\mathrm{O}$ (Output 2).

Step F2: Check—compute deviations and classify events

Given Input A (planned) and Input B (actual), the comparator computes deviations and determines event severity/class:

• Metric-level deviation rates. For each metric $j$, compute ${\delta }_j(t)$ using Equation (5).
• Aggregated deviation magnitude. Compute ${\Delta }_t$ using the weighted aggregation in Equation (6), consistent with the deployment weights $w_j$ defined during MOD.
• Adaptive thresholds and bands. Compute ${\mathsf{\Phi }}_j(t)$ using Equation (7) and construct trigger bands ${\mathsf{\Phi }}_{\mathrm{low}}(t)$ and ${\mathsf{\Phi }}_{\mathrm{high}}(t)$ using Equation (8).
• Event classification and response mode. Apply the prescriptive rule function $R\left({\Delta }_t\right)$ in Equation (9) to classify the event into Null, Self-Correction, or Escalation.

This procedure converts numeric deviations into a decision-relevant deviation event $e_t$, which contains the deviation magnitude, severity class, and context descriptors.

Step F3: Action—execute tiered response routines

FCA executes tiered responses consistent with Figure 2, using the responsibility–routine map (MOD Output 3) to determine who responds and which formal routine is activated.

Operational response (local corrective action). For contained deviations (typically Self-Correction, and some Escalation cases when local constraints are clear), operational actors execute predefined local adjustments and containment actions (e.g., SOP reinforcement, parameter tuning, short-cycle rework actions) and document execution evidence.

Control-layer response (root-cause analysis and action planning). For Self-Correction events, the control layer initiates a structured diagnosis routine, formulates an action plan, and coordinates closure evidence. This step prevents FCA from degenerating into ad hoc firefighting by ensuring that corrective actions are traceable and auditable.

Managerial response (resource reallocation; cross-unit coordination). For Escalation events, managerial routines are activated to address constraints that exceed local authority (e.g., staffing, maintenance windows, material supply coordination, scheduling trade-offs). Decisions are recorded as part of the event closure package.

Strategic response (target/threshold adjustment; policy update). When repeated or high-severity deviations indicate misfit between deployed targets and operating conditions, strategic routines may update objectives, thresholds, or governing policies. Such feed-forward updates are documented and linked to the specific event(s) that motivated the change.

4.3.3. Infrastructure and Support Module

MOD-FCA is designed as a prescriptive reference framework rather than an IT-specific solution. Nevertheless, consistent with Figure 2, effective instantiation typically requires two enabling components: a knowledge repository and a data/IT platform. These components provide the minimal infrastructure for traceability, routine execution, and continuous improvement.

Knowledge repository (organizational memory)

The knowledge repository stores the artifacts generated by MOD and FCA and enables reuse and learning across cycles. At minimum, it contains:

• Metric dictionary and linkage structure (MOD Output 1): definitions, units, sources, owners, and validated BM→VM linkages.
• Objective tensor and target profiles (MOD Output 2): $\mathrm{O}$ indexed by time–metric–scenario, including revision history.
• Responsibility–routine map (MOD Output 3): monitoring and response assignments, escalation paths, and routine templates.
• Threshold parameter history: ${\mathsf{\Phi }}_0$, ${\alpha }_1$, ${\alpha }_2$, $\gamma$, volatility references, and any revision rationale.
• Event logs: ${\Delta }_t$, ${\delta }_j\left(t\right)$, event class $R\left({\Delta }_t\right)$, and context descriptors.
• Closure records and improvement cases: root-cause analyses, actions taken, evidence of completion, effectiveness checks, and reusable best practices.

Functionally, the repository provides (i) traceability (why targets/thresholds changed), (ii) consistency (common semantics across levels), and (iii) learning (reuse of effective responses under similar contexts).

Data/IT platform (integration and connectivity)

The data/IT platform supports the capture, integration, and basic processing required to compute ${\mathrm{x}}_t$ (Equation (4)) and to automate event logging and routine execution. Typical capabilities include:

• Integration: connecting heterogeneous data sources (MES/ERP/QMS/sensors/manual inputs) into a unified data layer.
• Analytics and computation: computing ${\mathrm{x}}_t$, ${\delta }_j\left(t\right)$, ${\Delta }_t$, and threshold bands in a repeatable manner.
• Workflow enablement: issuing notifications, routing events to responsible roles, managing closure records, and maintaining audit trails.
• Access control and governance: ensuring data integrity, ownership clarity, and appropriate visibility across layers.

Importantly, MOD-FCA keeps the control logic explicit and auditable: digital tools enable execution, but the core comparator and response rules remain transparent (Equations (5)–(9)), thereby reducing the risk that control becomes embedded implicitly in opaque system behavior.

4.3.4. Parameter Initialization and Calibration

The metrics, weights, and threshold parameters in MOD-FCA are not intended to be universal constants. Instead, they are context-calibrated design parameters that should be initialized through a transparent and auditable procedure. First, candidate V-metrics are derived from stakeholder needs and prioritized through structured workshops or weighting methods such as AHP/weighted scoring. Candidate B-metrics are then generated from measurable deliverables and screened according to actionability, non-redundancy, and measurement stability. Second, deployment weights are assigned according to the strength of the linkage between B-metrics and V-metrics. Such linkages can be justified through 1:1 inheritance, 1:N decomposition, historical data analysis, or structured expert judgment when reliable data are not yet available. All weights are normalized and documented in the metric dictionary.

Third, the adaptive threshold parameters are initialized based on process standards, historical volatility, and strategic criticality. Specifically, Φ₀ represents the base tolerance level under normal operating conditions; α₁ controls how much the threshold expands in response to historical volatility; and α₂ controls how strongly the threshold is tightened for strategically critical metrics. The criticality parameter is assigned on a 0–1 scale based on the metric’s direct contribution to strategic value outcomes, quality/safety/sustainability implications, and managerial risk tolerance. During pilot operation, these parameters are reviewed against false alarms, missed deviations, and closure outcomes, and subsequent adjustments are recorded in the knowledge repository.

Table 2. Basis for initializing metrics, weights, and threshold parameters in MOD-FCA.

Element	Initialization Basis	Calibration/Documentation
V-metrics	Stakeholder needs, strategic objectives, power–interest assessment, structured workshops	Stored in the V-metric dictionary with definition, owner, unit, frequency, and data source
B-metrics	Deliverable decomposition, VQCECD coverage, actionability, non-redundancy, stable measurability	Stored in the B-metric dictionary and linked to responsible routines
BM→VM linkage	1:1 inheritance, 1:N decomposition, historical evidence, cross-role expert review	Retained only when semantically consistent and actionable
Deployment weights	Strength of causal/logical linkage, expert scoring, historical evidence where available	Normalized and updated when sufficient operational evidence accumulates
Φ0	Base tolerance from SOP standards, historical acceptable fluctuation, managerial risk appetite	Reviewed during pilot operation based on false alarms and missed deviations
α1	Sensitivity to historical volatility	Adjusted when normal process variation causes excessive false alarms
α2	Sensitivity to strategic criticality	Adjusted when critical deviations are under- or over-triggered
Criticality parameter	Directness to value outcomes, quality/safety/sustainability relevance, risk consequence	Recorded in the parameter profile and revised through feed-forward updates

summarizes the initialization basis and calibration documentation for the core elements in MOD-FCA. Therefore, the transferability of MOD-FCA lies in the parameter-setting procedure rather than in any fixed parameter value.

4.3.5. Operationalizing Sustainability Objectives Through MOD-FCA

Although MOD-FCA is not designed as a sustainability reporting framework, its operating logic can be used to embed sustainability objectives into daily management control. In the MOD module, sustainability concerns can first be formalized as value-centric metrics, such as energy intensity, carbon intensity, material utilization, waste/rework reduction, and resource efficiency. These value outcomes can then be decomposed into business-centric driver metrics, such as first-pass yield, rework rate, process stability, equipment efficiency, standard-work compliance, and closure cycle time. In this way, sustainability objectives are translated from high-level outcome indicators into controllable operational drivers.

In the FCA module, deviations in sustainability-related metrics can be monitored through the same plan–actual comparison and adaptive threshold logic used for quality, cost, and delivery metrics. Once deviations exceed predefined thresholds, corrective routines can be triggered and routed to responsible roles for local correction, cross-unit coordination, or target/threshold adjustment. Repeated sustainability-related deviations and effective responses can also be codified in the knowledge repository as reusable improvement cases. Therefore, the sustainability relevance of MOD-FCA lies in transforming sustainability objectives from ex post reporting items into actionable control objects that can be deployed, monitored, corrected, and learned from within daily operations.

4.4. Design Principles

Based on the construction and instantiation of MOD-FCA, this study distills four transferable design principles. Following the contribution typology of Gregor and Hevner [28], these principles constitute Level 2 prescriptive design knowledge (nascent design theory) for digitally enabled management control. They aim to guide organizations operating in data-intensive environments to build closed-loop control arrangements that are computable, traceable, and executable.

Table 3. Mapping between MOD-FCA design principles and framework components.

Design Principle	Corresponding MOD-FCA Components	Main Mechanism
Metrics-driven formalization	V-metric dictionary; B-metric dictionary; BM→VM linkage structure; objective tensor	Translates strategic intent into measurable, time- and scenario-indexed control objects
Process-embedded accountability	Responsibility–routine map; metric owners; monitoring/responding roles; escalation paths	Binds metrics and deviation classes to responsible roles and formal routines
Algorithmic closed-loop governance	Performance vector; deviation computation; adaptive thresholds; trigger bands; event classification rules	Converts plan–actual gaps into classified deviation events and tiered responses
Evolutionary knowledge accumulation	Event logs; closure records; improvement cases; knowledge repository; feed-forward updates	Converts repeated deviation handling into reusable organizational knowledge and parameter/routine refinement

summarizes how each principle is operationalized in MOD-FCA.

Principle 1: Metrics-driven formalization

Traditional control practices often compress strategic intent into a small set of financial scalars. Literature has long criticized such reliance on lagging financial indicators for weakening the traceable links between value objectives and frontline behaviors, and for failing to accommodate temporal variation in dynamic environments [19,20]. To address this, MOD-FCA requires that stakeholder needs and strategic intent be formalized as a multi-dimensional value-metric vector, and that targets be represented through an objective tensor that captures time and scenario differences. The purpose of using vector and tensor representations is not to increase mathematical complexity, but to provide a unified grammar for deployment mapping and plan–actual comparison. This formalized structure ensures vertical consistency between strategic intent and operational metrics, reducing the risk that local optimization erodes overall value realization [34,50].

Operationalization in MOD-FCA: V-metrics dictionary; objective tensor $\mathrm{O}(T\times N\times S)$; linkage structure for deployment (MOD outputs).

Principle 2: Process-embedded accountability

Digital transformation often dramatically increases the number of monitorable items, yet response ownership and decision rights are frequently not clarified accordingly. This results in a phenomenon of information waste, where organizations can see deviations but lack the mechanism to trigger effective action [12,14]. MOD-FCA embeds control points directly into routines and standardized work by explicitly defining who monitors, who responds, and when to escalate within specific SOPs. A role–responsibility structure is used to codify metric owners and the boundaries of corrective authority [30]. Accountability alignment is a prerequisite for a functioning closed loop: when responsibility scope and control scope are misaligned, deviation signals are dissipated by organizational frictions and become noise rather than actionable control information.

Operationalization in MOD-FCA: responsibility–routine map (who monitors/responds/escalates); routine templates and escalation paths (MOD Output 3).

Principle 3: Algorithmic closed-loop governance

Under high-frequency data streams, purely manual monitoring can lead to cognitive overload, delayed responses, and subjective judgment bias [10,41]. MOD-FCA emphasizes that closed-loop control must be computable: deviations are translated into classifiable events through an adaptive threshold function $\mathsf{\Phi }\left(t\right)$, and then used to trigger tiered response routines. Routine fluctuations are handled locally (e.g., automated tickets/alerts/short-cycle adjustments), whereas structural deviations trigger escalation and parameter updates (e.g., adjustments to thresholds, targets, or resource allocations), combining feedback and feed-forward governance [26]. The principle aims to reduce managerial noise caused by false alarms and missed detections, and to shorten the latency from deviation occurrence to action execution [12].

Operationalization in MOD-FCA: deviation computation ${\delta }_j\left(t\right)$, aggregated deviation ${\Delta }_t$, adaptive thresholds $\mathsf{\Phi }\left(t\right)$, and event classification/response rules (Equations (5)–(9)); tiered responses across layers (Figure 2).

Principle 4: Evolutionary knowledge accumulation

If corrective actions remain one-off interventions, organizations repeatedly pay the learning cost for similar problems and struggle to build a stable control capability. MOD-FCA requires that closure outcomes be codified into structured knowledge of “deviation–response–effectiveness” and stored in a knowledge repository. These cases then inform subsequent threshold setting, routine refinement, and target deployment adjustments. This mechanism enables not only corrective actions under existing rules (single-loop learning), but also evidence-based revision of rules and target-setting logic when warranted (double-loop learning) [22]. As operations accumulate, the reusability and adaptability of control logic improve, and the closed loop gradually shifts from reliance on individual experience to reliance on organizational memory.

Operationalization in MOD-FCA: event log + closure records; knowledge repository (rules/cases/best practices); feed-forward updates to targets/thresholds/policies (Figure 2).

5. Evaluation

This section aims to empirically evaluate the MOD-FCA framework through case studies. Following the design science research paradigm, we conducted multiple rounds of exploratory case studies with five enterprises from the casting, energy, automotive finance, engine manufacturing, and small appliance manufacturing industries during the framework development phase. This process continuously optimized the framework’s generalizability and component design (i.e., formative evaluation). On this basis, we selected the most representative Casting company (Enterprise A) as an in-depth longitudinal case (i.e., summative evaluation) to demonstrate the implementation process of MOD-FCA in detail and examine its instantiation, process evidence, and perceived usefulness [49]. It should be noted that the evaluation does not claim to provide a causal before-and-after test of operational performance. During the study period, Enterprise A’s digital-intelligent platform and standardized KPI capture mechanism were still under construction, and reliable, comparable longitudinal KPI series before and after implementation were not available. Therefore, the present evaluation focuses on perceived usefulness and traceable process evidence, including interviews, workshop records, internal documents, and implementation artifacts [24]. Therefore, the present evaluation should be interpreted as an instantiation-based assessment of applicability, operability, and perceived usefulness. It does not support statistical generalization to all companies, nor does it constitute a hard confirmation of actual operational effectiveness.

To support the evaluation, we relied on multiple complementary sources of evidence collected during the instantiation. These include: (i) semi-structured interviews with managerial and operational roles to capture governance pain points and perceived usefulness; (ii) workshop and meeting records documenting design decisions, iterations, and agreement on metric semantics and routines; (iii) internal documents such as SOPs, process descriptions, and performance reports used to reconstruct baseline control practices and verify changes; and (iv) implementation artifacts produced by the instantiation (e.g., metric dictionaries, objective profiles, responsibility–routine mappings, and deviation/closure records), which provide traceable process evidence. These sources were triangulated to corroborate key claims about multi-layered traceability, deviation actionability, and knowledge codification.

5.1. Background

5.1.1. The Enterprise Profile and Industry Dilemma

Enterprise A is a large state-owned automotive component casting enterprise with approximately 5000 employees. It operates in a hierarchical manufacturing environment characterized by multiple workshops, cross-functional coordination requirements, stringent quality-control demands, and relatively standardized but still unevenly executed production routines. Before the project, the enterprise had introduced several digital systems and had begun its digital-intelligent transformation, but its data foundation remained fragmented: key operational records were partly system-based and partly manual, data definitions were not fully unified across units, and many control routines still depended on periodic reports and managerial experience.

These characteristics made Enterprise A a suitable but demanding setting for evaluating the instantiation of MOD-FCA. On the one hand, its hierarchical structure, measurable production processes, and strong need for cross-level coordination created clear conditions for applying a multi-layered closed-loop control framework. On the other hand, its weak data standardization and uneven process discipline reflected the typical implementation frictions faced by traditional manufacturing enterprises during digital transformation. Therefore, the case provides a stringent context for examining the operability and perceived usefulness of MOD-FCA, but it does not imply automatic generalizability to all organizational forms or industries.

5.1.2. Initial Diagnostic: Recurring Governance Gaps

Early field investigation suggested several recurring governance gaps that shaped the instantiation focus on Enterprise A. First, although many indicators existed, outcome fluctuations were often difficult to trace back to a small set of controllable process drivers, and indicator definitions were not always consistent across units. Second, deviation handling relied heavily on periodic reporting and ad hoc interventions; escalation and closure routines were unevenly institutionalized, with occasional decoupling between reported compliance and actual execution. Third, improvement knowledge was largely person-dependent and weakly codified, leading to recurring problems across workshops and shifts. These observations served as a problem-structuring lens for configuring metrics, targets, routines, and triggering rules in the subsequent MOD-FCA instantiation; they do not presuppose the effectiveness of the framework, which is evaluated longitudinally in the following sections.

5.2. MOD-FCA Implementation Process

Enterprise A initiated a business-architecture-oriented management redesign program in May 2024. Given the novelty of the initiative and the lack of specialized internal resources, the enterprise engaged the research team (the authors) to support its implementation. The instantiation lasted eight months (May–December 2024) and involved 15 core business units. Consistent with an iterative field-based engagement, the implementation progressed through three stages: combined diagnosis, design, and organizational adoption.

Stage 1: Joint diagnosis and business architecture articulation.

A joint working group was established with representatives from the major business units. Through facilitated workshops and field investigation, the team clarified the end-to-end operational logic—from raw materials to qualified castings—and articulated the key value-creating activities, control points, and information objects. This stage transformed fragmented tacit knowledge into an explicit business architecture that provided a common basis for subsequent metric design and control deployment.

Stage 2: Framework instantiation and mechanism design.

Building on the articulated architecture, MOD-FCA was instantiated into the enterprise context by designing (i) a vertically aligned metric system and associated definitions, (ii) objective profiles for key metrics to support consistent target deployment and monitoring, and (iii) deviation-handling routines that link plan–actual comparison to corrective actions and escalation. In parallel, the team initiated the codification mechanism by structuring how deviations, responses, and closure evidence would be documented and accumulated as reusable organizational knowledge. Figure 4 illustrates the blueprinting logic through which Enterprise A configured a multi-layered closed-loop management control architecture under MOD-FCA guidance. The blueprint defines cross-level roles and responsibilities, the planned control baseline (metrics and objective profiles), and the feedback-driven deviation governance mechanism (event classification, tiered response routines, closure discipline, and knowledge codification).

Stage 3: Piloting, tuning, and routine stabilization.

The framework then entered a period of pilot operation and refinement across the participating units. Key parameters (e.g., indicator weights and control thresholds) were tuned based on early operational feedback, and response routines were reinforced to ensure timely handling and verifiable closure of deviations. Over time, closure records and improvement cases accumulated, supporting continuous improvement and enhancing the stability and reusability of the control logic.

Overall, the three-stage process enabled Enterprise A to move from fragmented measurement and ad hoc interventions toward a more structured, closed-loop control arrangement that is traceable across levels and increasingly supported by codified learning.

5.3. Case Instantiation in Standardized Work Management

This section uses the Standardized Work Management domain within the Production Management function as an illustrative instantiation to demonstrate the end-to-end process of dual-perspective metric design, formal modeling, and closed-loop governance under MOD-FCA.

5.3.1. Value-Centric Metrics Design: From Stakeholder Needs to Strategic Outcomes

Following the value-centric metric design procedure described in Section 4.3.1 (A1–A5), we defined the strategic value orientation for the Production Management function by translating stakeholder needs into measurable value-centric metrics (V-metrics) (partial results are shown in

Table 4. Design of value-centric metrics addressing stakeholders’ needs satisfaction.

Stakeholders	Needs	Metrics	Needs Prioritization	Value-Centric Metrics(VC)
Shareholders	Implement parent company mandates in production operations;	Mandate compliance rate.	1/14	Mandate compliance rate; Product mix fulfillment rate, First-time acceptance rate, On-time delivery rate, Capacity utilization; OEE, Inventory turnover ratio; Conformance rate, First-pass yield, Cpk; Unit production cost, Controllable cost per unit; Energy intensity per RMB 10,000 output, Carbon intensity per RMB 10,000 output, Specific energy consumption; Lean maturity index; Online transition rate, Model utilization rate, AI integration rate; Competency development completion rate, Role competency rate; Order fulfillment rate (with loss rate tracking); On-time shipment rate, Order accuracy rate; Preservation compliance rate; Ramp-up cycle time; Capacity utilization analysis accuracy;
	Maintain controlled processes to fulfill customer specifications (volume/quality/delivery);	Product mix fulfillment rate, First-time acceptance rate, On-time delivery rate, Capacity utilization, Production schedule adherence.	2/14
	Drive continuous productivity improvement;	Efficiency target achievement rate, Overall Equipment Effectiveness (OEE), Inventory turnover ratio, Modernization project completion.	3/14
	Enhance product quality through systematic controls;	Conformance rate, First-pass yield, Process capability index (Cpk), Rework volume.	4/14
	Optimize production cost structure;	Unit production cost, Controllable cost per unit (tooling/mold/maintenance/ packaging/logistics), Outsourcing fees.	5/14
	Advance energy efficiency management systems;	Energy intensity per RMB 10,000 output, Carbon intensity per RMB 10,000 output, Specific energy consumption.	6/14
	Establish a lean manufacturing framework;	Lean maturity index, 5S compliance rate, Benchmark teams established.	7/14
	Accelerate digital-physical integration in production;	Online transition rate, Model utilization rate, AI integration rate.	8/14
	Develop high-performance production leadership.	Competency development completion rate, Role competency rate.	9/14
Marketing department	Ensure maximum responsiveness to customer orders;	Order fulfillment rate (with loss rate tracking), Order review cycle time, Full-order completeness.	10/14
	Execute precision shipment operations;	On-time shipment rate, Order accuracy rate.	11/14
	Guarantee product integrity through delivery.	Customer acceptance rate, Preservation compliance rate.	12/14
R&D department	Achieve rapid production ramp-up for new products;	Ramp-up cycle time.	13/14
	Implement data-driven capacity planning;	Capacity utilization analysis accuracy,	14/14
etc.	etc.	etc.	etc.	etc.

). We identified key stakeholders (including shareholders, the Marketing Department, and the Product R&D Department) and elicited their core expectations regarding production management. Qualitative needs were converted into quantifiable indicators, and priorities were determined through a power–interest assessment and structured workshops.

Focusing on high-priority needs, we refined a compact set of V-metrics as the value orientation for the domain. In this case, VC1 (Product First-Pass Yield) was selected as the primary value objective that the Standardized Work Management domain is expected to support. The challenge target for VC1 was set to increase from 81.2% to 86.5%.

In addition to quality and delivery-oriented indicators, the value-centric metric set also included sustainability-related indicators such as energy intensity per RMB 10,000 output, carbon intensity per RMB 10,000 output, and specific energy consumption. These indicators were not treated as isolated environmental reports. Instead, under MOD-FCA, they can be connected to controllable business-centric drivers such as first-pass yield, rework reduction, process stability, equipment efficiency, and closure cycle time. In this way, sustainability-related outcomes are embedded into the same deployment and feedback logic as operational performance indicators.

5.3.2. Architectural Deconstruction: Deliverables and Business-Centric Metrics

To operationalize VC1, we conducted an architecture-oriented deconstruction of the Standardized Work Management domain, identified its key deliverables, and designed business-centric metrics (B-metrics) accordingly. This step establishes the driver layer required for traceability and actionability.

(1) Deliverables as measurable control objects

To operationalize standardized work governance, we decomposed the domain into five measurable deliverables (artifacts/evidence) that can be monitored and managed across organizational layers:

D1: SOP standard artifacts (approved SOP versions and their availability).

D2: SOP execution evidence (compliance observations and execution logs).

D3: SOP design governance artifacts (templates and change-control/revision records).

D4: SOP competence artifacts (training, certification, and assessment records).

D5: SOP assurance artifacts (inspection/audit reports, nonconformance records, and verified closure evidence).

Collectively, these deliverables represent the governance chain through which standardized work is designed, deployed, executed, and verified—thereby providing actionable levers that can influence VC1.

(2) VQCECD-guided metric tailoring

VQCECD was used as a coverage discipline to generate candidate driver metrics. However, metrics were tailored to ensure (i) actionability (clear ownership and routines), (ii) non-redundancy (avoiding dual proxies for the same lever, e.g., labor-hours and cycle time simultaneously), and (iii) stable measurability in the pilot context. In this instantiation, Compliance is scoped to SOP procedural compliance (adherence to defined steps and quality-critical points), rather than broader regulatory compliance.

Table 5. Business-centric metrics for standardized work governance (selected set).

Code	Deliverable	VQCECD	Metric Definition	Primary Linkage to VC1
BC1	D1	Volume	SOP coverage rate: % of key operations with an approved, current SOP accessible at point of use	Indirect (enables BC5)
BC2	D1/D3	Quality	SOP document quality pass rate: % SOPs passing completeness/clarity review checklist	Indirect (enables BC5)
BC3	D3	Efficiency	SOP change lead time (median days): change request → approved SOP version release	Indirect (prevents mismatch-driven defects)
BC4	D4	Volume	Training/certification coverage: % operators certified for the SOP set relevant to their station	Indirect (enables BC5)
BC5	D2	Compliance	Execution compliance rate: % observations meeting all quality-critical SOP steps	Direct driver
BC6	D2	Efficiency	Standard work stability: % cycles within standard cycle-time tolerance band (or cycle-time variance index)	Indirect (process stability → quality)
BC7	D5	Quality	Repeat deviation rate: % SOP-related deviations recurring within a defined window	Direct/indirect (sustained correction)
BC8	D5	Cost	Monitoring effort: inspection/audit labor-hours per 100 operations (or per shift)	Indirect (cost of control; complements proactive feedback)
BC9	D5	Digital	Digital traceability rate: % SOP/training/inspection records captured with traceable IDs in the system	Indirect (timeliness & credibility of feedback)
BC10	D5	Efficiency	Closure cycle time (median days): deviation identified → verified closure evidence	Direct/indirect (prevents accumulation & rework)

reports a compact set of B-metrics selected for reporting and governance routines. The full metric dictionary contained additional local metrics, but the listed metrics represent the primary controllable drivers with the strongest link to the value-centric outcome VC1 (First-Pass Yield).

(3) BM→VM linkage and verification

Following the linkage verification gate in Section 4.3.1 (C1–C2), candidate BC metrics were screened through cross-role workshops and review of historical deviation records to retain only linkages that were (i) semantically consistent across units and (ii) actionable through defined routines. In the finalized linkage logic, BC5 is treated as the primary direct driver for VC1, while BC1/BC2/BC4 provide readiness conditions, and BC7/BC10 capture the effectiveness and timeliness of closed-loop correction that stabilizes improvements over time.

5.3.3. Multi-Scenario Target Setting and Quantitative Deviation Perception

Objective tensor. Following Equation (3), we configured objective profiles across multiple time periods, multiple indicators, and multiple scenarios, and instantiated them as an objective tensor $\mathrm{O}(T\times N\times S)$. Here, $T$ was defined on a monthly cadence, $N=10$ (BC1–BC10), and $S$ included baseline and challenge scenarios. For example, the challenge target for BC5 was set to 98%.

Deviation quantification. During operation, actual performance is mapped into the performance vector ${\mathrm{x}}_t$ (Equation (4)). In Enterprise A, a pilot AI-enabled visual monitoring module supported the collection of selected shop floor compliance signals. Deviations are then computed using the metric-level deviation rate ${\delta }_j\left(t\right)$ (Equation (5)) and aggregated into the comprehensive deviation magnitude ${\Delta }_t$(Equation (6)), forming a quantitative basis for subsequent event classification and response triggering.

5.3.4. Digitally Enabled Closed-Loop Control

Adaptive thresholding. We configured the adaptive threshold function ${\Phi }_j\left(t\right)$ following Equation (7), where thresholds vary with historical volatility and strategic criticality. In this domain, BC5 was treated as strategically critical (criticality parameter set to 0.9 in the configured profile), resulting in tighter triggering sensitivity for compliance deviations. This value does not represent a universal coefficient, but a case-specific setting derived from the parameter initialization procedure described in Section 4.3.4. The rationale is threefold. First, BC5 measures execution compliance with quality-critical SOP steps and was identified as the most direct controllable driver of VC1 (Product First-Pass Yield). Second, deviations in BC5 may immediately generate quality defects, rework, and process instability, making delayed responses costly. Third, cross-role workshops with production, quality, and management personnel classified BC5 as a high-priority control metric relative to other readiness or support metrics such as SOP coverage and training coverage. Therefore, the value of 0.9 was used to increase triggering sensitivity for compliance deviations during the pilot, while subsequent adjustments were designed to be documented in the knowledge repository. Trigger bands ${\Phi }_{\mathrm{low}}\left(t\right)$ and ${\Phi }_{\mathrm{high}}\left(t\right)$ were constructed following Equation (8).

Furthermore, because the study focused on early-stage framework instantiation and the enterprise’s digital-intelligent platform was still under construction, a full statistical sensitivity analysis was not conducted. Instead, the pilot relied on scenario-based parameter calibration and managerial review, in which threshold settings were examined against false alarms, missed deviations, and closure feasibility. Future research should systematically conduct formal sensitivity analyses once richer longitudinal data become available.

Tiered triggering and correction. Based on the deviation magnitude ${\Delta }_t$ and the trigger bands, the response mode was determined using the rule function $R\left({\Delta }_t\right)$(Equation (9)). In practice, when the deviation pattern entered the Self-Correction range, the system initiated predefined routines (e.g., operational reminders and short-cycle checks) and documented closure evidence. When deviations entered the Escalation range, the workflow triggered formal governance routines, including cross-level coordination and, when warranted, structured revision of SOP documentation and related routines. In this manner, deviation governance moved beyond periodic reporting toward a closed-loop arrangement with explicit triggering, accountability, and closure discipline.

5.3.5. Knowledge Evolution and Organizational Learning

All deviation-handling cases were documented in a structured format and stored in the knowledge repository as reusable “deviation–response–outcome” patterns. Beyond accumulation, the repository was subject to periodic knowledge governance (e.g., consolidation of duplicates, refinement of classification, and updating of recommended actions), supporting higher-quality organizational learning. Over time, this mechanism helps transform individual experience into organizational assets and enables disciplined feed-forward updates (targets, thresholds, and routines) when repeated deviations indicate misfit between objectives and operating conditions.

5.4. Qualitative Feedback

To complement the process evidence from the instantiation, we collected qualitative feedback from multiple organizational roles after the framework was put into pilot use. This approach enables a triangulated assessment of perceived usefulness, enhancing the validity of the case findings [49,52]. Qualitative inputs were obtained through semi-structured interviews with senior managers and department heads conducted at different project stages. Interview materials were synthesized via thematic consolidation to identify recurring perceptions regarding governance improvements, remaining frictions, and refinement opportunities.

Across interviews, respondents consistently described that MOD-FCA helped address several persistent governance dysfunctions that had constrained performance and transformation readiness. These perceptions align with the blueprint logic summarized in Figure 4—namely, clarifying the planned control baseline (metrics/objectives), enabling deviation visibility, and institutionalizing tiered responses. The feedback was categorized into the following key themes.

Theme 1: From completion-rate reporting to a traceable, multi-layered metric system (problem visibility and diagnosis). Respondents emphasized that pre-project measurement practices were dominated by completion-rate indicators and lagging financial results. Literature suggests that such rear-view metrics often fail to reveal whether business capabilities are actually improving [19,20]. After the implementation of MOD-FCA, the dual-perspective metric structure (value-centric outcomes linked to business-centric drivers) significantly improved diagnostic traceability. When outcomes deteriorated, managers were better able to identify controllable driver metrics and the responsible routines. One executive summarized this shift as moving from meeting assessment items to seeing what really drives performance. This corroborates the view that effective PMS must facilitate the causal understanding of performance drivers rather than merely monitoring outputs [14].

Theme 2: From information distortion and problem hiding to greater process transparency (truthful signals). Multiple respondents noted that frontline staff previously tended to avoid exposing problems due to a fear of blame, leading to data being “smoothed” before reaching top management. This phenomenon is consistent with the gaming and data manipulation behaviors often observed in directive, pressure-intensive control environments [39,53]. MOD-FCA was perceived to improve transparency by (i) shifting the control focus toward process-embedded driver metrics and (ii) emphasizing deviation events and closure evidence rather than narrative reporting. A department head described that, previously, “data could not truly reflect the situation,” whereas after the pilot, deviations were more likely to be discussed as objects for resolution rather than as triggers for blame. This suggests that MOD-FCA fosters an enabling rather than coercive use of control information, encouraging truthful reporting and problem-solving [54,55].

Theme 3: From leadership inaction or frontline firefighting to role-aligned accountability and collaboration (execution assurance). A recurring pre-project issue was the mismatch between authority and responsibility: leaders either remained detached from operational problems or were forced to engage in frontline “firefighting.” Theoretically, this reflects a misalignment between the span of control (resources available) and the span of accountability (outcomes responsible for), a disconnect known to hamper strategic execution [56,57]. Interviewees attributed improvements to clearer role boundaries and routine bindings facilitated by MOD-FCA. By specifying who monitors, who responds, and when escalation occurs (as depicted in Figure 4), the framework reduced the ambiguity that often leads to role stress and inaction [58].

Theme 4: From inspection-driven correction to process-driven closure discipline (closed-loop governance). Respondents highlighted that even when problems were exposed, governance often depended on repeated ad hoc checks rather than a process-driven closed loop. MOD-FCA was perceived to strengthen closure discipline by linking deviation detection to predefined routines and requiring closure evidence. This marks a shift from a passive PMS to an active PMS, where information is used not just for monitoring but to trigger and verify corrective action [20,33]. This reduced the reliance on personal supervision as the primary control mechanism and shifted governance toward a routine-based, auditable cycle (deviation → response → closure → learning), thereby mitigating the risks of gaming or superficial compliance often found in low-trust environments [39].

Taken together, the qualitative feedback suggests that MOD-FCA was not perceived merely as another reporting system, but as a governance logic that improves: (i) the visibility of actionable problems, (ii) the accuracy and credibility of control information, (iii) role-aligned accountability, and (iv) routine-based closure. By addressing these dimensions, MOD-FCA reduces the transaction costs associated with goal achievement under ad hoc interventions and moves the organization toward what Ferreira and Otley [22] described as a coherent system where control elements work in concert rather than in isolation.

5.5. Discussion

This longitudinal case provides process evidence for the operability and perceived usefulness of MOD-FCA in addressing persistent governance problems in traditional management control settings. In Enterprise A, pre-existing dysfunctions—such as leadership being forced into frontline firefighting, frontline problem hiding, and metric systems dominated by lagging completion rates—constrained the organization’s ability to ensure that resources were obtained and used effectively [21]. These issues reflect the classic loose coupling between strategic intent and operational reality, often criticized in MCS literature [23]. By configuring a multi-layered control blueprint (Figure 4) and operationalizing a closed-loop logic, MOD-FCA helped shift control from episodic supervision to traceable routines, and from outcome-only reporting to actionable driver-based governance. Below, we discuss the theoretical and practical implications of these findings.

5.5.1. Theoretical Implications

First, MOD-FCA contributes to the MCS/PMS literature by providing prescriptive, mechanism-oriented design knowledge that complements established descriptive frameworks. While classic frameworks clarify “what” elements constitute MCS (e.g., objectives, measures, rewards) [22,33], they have been critiqued for functioning primarily as heuristic checklists while offering limited guidance on “how” to configure traceable cross-level linkages [35]. MOD-FCA addresses this meso-level design gap by specifying a deployable structure: it links value-centric outcomes to business-centric drivers and couples them with objective profiles and responsibility–routine mappings. In this case, decomposing a strategic value objective (e.g., VC1) into controllable process drivers (e.g., BC5) provided an operational pathway that strengthens the actionability of diagnostic control, moving beyond the static monitoring often associated with traditional systems [18,20].

Second, this study illustrates how cybernetic control logic can be upgraded for digitally enabled operations through computable deviation governance and tiered responses. Rather than treating control as periodic variance reporting, MOD-FCA formalizes plan–actual comparison into deviation events and associates them with tiered routines (self-correction vs. escalation). Theoretically, this extends the cybernetic view by adopting a SoS perspective, where autonomous control loops at the operational level are interconnected with managerial and strategic loops to handle complexity [27]. By specifying how feedback signals are translated into role-aligned responses, the framework reduces the decision latency and information waste that often plague data-rich environments [12].

Third, MOD-FCA highlights the interplay between control and learning by making knowledge codification an integral part of closed-loop governance. The case indicates that recurring issues were previously handled in an episodic manner and remained person-dependent; MOD-FCA supports systematic accumulation of “deviation–response–outcome” cases and disciplined feed-forward adjustments of targets, thresholds, and routines. This provides a concrete mechanism through which control systems can enable continuous improvement and organizational learning in data-intensive environments.

5.5.2. Practical Implications

For practitioners, the primary value of MOD-FCA is its role as a blueprint for management logic that can precede and guide digitalization investments. In many legacy enterprises, transformation efforts risk becoming technology accumulation without a stable governance design. The Enterprise A case suggests that MOD-FCA can reduce this risk by clarifying: (i) what should be measured (driver–outcome metrics), (ii) what objectives and tolerances should be deployed (objective profiles and threshold bands), (iii) who should respond and how (role-aligned routines and escalation paths), and (iv) how to ensure closure and learning (closure records and knowledge repository). These mechanisms directly address common pain points observed in practice: leadership over-involvement in frontline firefighting, unreliable upward information flows caused by problem hiding, metric systems that cannot reveal or trace root causes, and a lack of process-driven closure.

At the same time, the case indicates boundary conditions for application. The current evidence suggests that MOD-FCA is most applicable to organizations that meet three basic conditions. First, the organization should have relatively identifiable value outcomes and controllable process drivers, so that V-metrics and B-metrics can be meaningfully linked. Second, the organization should possess or be willing to build a minimum data foundation, including stable metric definitions, traceable records, and basic workflow documentation. Third, the organization should have a sufficient hierarchical or role-based governance structure to support differentiated responsibilities for monitoring, diagnosis, escalation, and closure. Accordingly, MOD-FCA may be more directly transferable to manufacturing and operations-intensive contexts with measurable processes, recurring routines, and cross-level coordination needs. Its application in flatter organizations, highly creative work, project-based services, or digitally immature firms would require adaptation. In such contexts, the framework should be introduced gradually, beginning with a focused pilot domain where process boundaries, data sources, and accountability relationships are relatively clear.

From a sustainability perspective, MOD-FCA provides a way to integrate operational sustainability into daily management control rather than leaving it as a periodic reporting exercise. For manufacturing enterprises, sustainability outcomes such as energy intensity, carbon intensity, material waste, rework, and resource utilization are often shaped by routine operational decisions. MOD-FCA helps translate these outcomes into controllable process drivers, assigns monitoring and response responsibilities, and triggers corrective routines when deviations occur. This is particularly relevant for traditional manufacturing firms because many sustainability losses arise from process instability, repeated deviations, delayed closure, and fragmented accountability. By strengthening traceability and closure discipline, MOD-FCA can support sustainability-oriented continuous improvement, although the long-term sustainability effects still need to be tested with longitudinal quantitative indicators.

6. Conclusions, Limitations, and Future Research

Facing increasingly volatile and data-intensive operating conditions, enterprises need management control that not only measures performance but also enables timely correction and disciplined learning. This study addressed the practical and theoretical challenge of translating data abundance into actionable governance by developing MOD-FCA: a quantitative reference framework for multi-layered closed-loop management control in the digital era.

This research contributes to the literature in three specific ways. First, it advances MCS/PMS research by offering prescriptive design knowledge that complements established descriptive frameworks. Unlike taxonomic models that list what to control, MOD-FCA specifies a deployable structure that links value-centric outcomes to business-centric driver metrics. It connects metrics to objectives, routines, and accountability in a traceable manner, constituting a Level 2 contribution to the design science knowledge base. Second, it formalizes a computable closed-loop control logic. By operationalizing plan–actual comparison into deviation events and tiered response routines, the framework strengthens actionability beyond periodic reporting. This responds to the call for MCS to evolve from static monitoring to dynamic, predictive governance, reducing the decision latency often observed in traditional systems. Third, it incorporates knowledge codification as an integral component of closed-loop governance. By supporting the accumulation of deviation–response–outcome cases, the framework institutionalizes double-loop learning capabilities. This ensures that control activities drive continuous improvement rather than merely enforcing compliance.

From a managerial perspective, the study provides a practical pathway for organizations pursuing digital transformation. Consistent with the “Lean first, then Digitalize” principle, the study suggests that enterprises should first establish an explicit control blueprint—clarifying what to measure, how to deploy targets, how to trigger responses, and who is responsible. This logic-first approach helps optimize information flow and reduce digital waste, thereby improving business-IT alignment and reducing reliance on ad hoc supervision and frontline firefighting.

This study also has limitations. Although this study provides longitudinal process evidence and qualitative feedback from multiple organizational roles, it does not include systematic pre- and post-implementation KPI comparisons. Therefore, the current findings should be interpreted as evidence of design utility, operability, and perceived usefulness, rather than as conclusive evidence of actual performance effectiveness. Future research should collect standardized longitudinal indicators, such as deviation response time, closure cycle time, target achievement rate, repeat deviation rate, inspection workload, and resource/energy consumption indicators, to examine the long-term operational and sustainability effects of MOD-FCA through multi-case or quasi-experimental designs. The framework was validated primarily in hierarchical manufacturing settings and through an in-depth instantiation in a single focal enterprise; therefore, transferability to other organizational forms (e.g., flat organizations or service contexts) warrants further examination. Additionally, future research could extend the evaluation with richer longitudinal data to assess the long-term impact of the framework on organizational culture and strategic agility.

Finally, the implementation of MOD-FCA may encounter practical barriers, particularly in organizations with low digital maturity. Poor data quality, inconsistent metric definitions, fragmented information systems, and weak process standardization may increase the effort required to build reliable performance vectors and threshold rules. In addition, employees may resist more transparent deviation tracking if the system is perceived as a blame-oriented monitoring tool rather than an enabling mechanism for problem solving. Therefore, successful implementation requires not only technical configuration but also data governance, role clarification, training, and change-management efforts. Future research should examine the adoption costs and organizational change mechanisms associated with MOD-FCA implementation in digitally immature enterprises.

In summary, MOD-FCA enriches the MCS field by providing a theoretically grounded and practically instantiated reference framework. It offers a structured approach to making metrics actionable, responses executable, and learning institutionalized—capabilities that are increasingly essential for operational governance in the digital era.