Reproducibility Standards for Lean Maturity Models: Design Guidelines for Logistics Operations

Abstract

Background: Lean management has been widely adopted, particularly in logistics operations. Achieving lean is not a discrete intervention but a continuous process of maturing in the integration of processes, work systems, and organizational capabilities within a coherent management philosophy. This maturation requires structured measurement instruments for tracking maturity progression. Although numerous lean maturity models (LMMs) have been proposed, none has achieved methodological standardization or acceptance as a measurement yardstick. Methods: This study addresses this gap by evaluating 27 qualified LMMs using a reproducibility-inspired assessment framework. The paper introduces the OVRGP framework (Opportunity, Validity, Reliability, Generalizability, and Process Integrity), comprising 17 rigor-based criteria. Independent raters with substantial lean expertise evaluated all 27 models using a six-point ordinal scale, achieving a pre-consensus inter-rater reliability of ICC(2,1) = 0.836. Results: Nine critical methodological weaknesses were identified, with average scores below 2.0 for criteria requiring empirical validation, structural integrity testing, and cross-context replication. Conclusions: The study offers targeted methodological guidelines for strengthening future LMM development in logistics and supply chain contexts, and introduces the OVRGP framework as a universal reference architecture for maturity model development across industries. It provides researchers, organizations, and consulting practitioners with a design reference standard for rigorous lean maturity instruments.

1. Introduction

Since its origins within Toyota in the 1950s, the Toyota Production System (TPS) has evolved into one of the most influential paradigms in operations management [1]. Its international diffusion accelerated following The Machine That Changed the World [2], which popularized the term “lean” [3]. Since then, lean has been embraced by almost every industry, and successful adaptation has given enterprises significant performance advantages [4,5]. However, sustained performance improvements depend not on isolated tool implementation but on systemic integration of processes, information flows, organizational routines, and leadership practices [6,7,8]. Structured instruments to assess the level of lean maturity are therefore essential. The literature is characterized by measures attempting to assess the impact of lean, yet most exhibit a dominant bias towards financial benefits [9,10], which are ill-equipped to measure constructs such as lean integration [11].

Recognizing this inherent research gap, many studies have attempted to measure the degree of lean integration, using terms such as leanness, lean adaptation level, and lean maturity level. These attempts are well documented by comprehensive systematic reviews [12,13,14]. Numerous lean maturity models (LMMs) have therefore been developed, yet despite their proliferation, no LMM has emerged as a methodological benchmark in academia or practice. While prior reviews have cataloged model content and structural configurations, little attention has been given to evaluating their methodological rigor, and no study has examined whether existing LMMs are fit-for-purpose as scientific measurement instruments. This gap is further compounded in logistics and supply chain contexts, where recent research increasingly frames performance as the result of integrated capability configurations spanning facilities and partners, suggesting that lean maturity assessment should evaluate systemic capability rather than isolated tools or short-term financial impacts [15,16,17,18]. Yet despite this strategic importance, no validated lean maturity model exists specifically designed for logistics and supply chain as an integrated organizational capability [19,20], and maturity models in this domain broadly suffer from limited validation and the absence of improvement guidelines [21].

This dual absence of a rigorous LMM evaluation framework generally, and of targeted methodological guidance for logistics and supply chain contexts specifically, represents the central research gap motivating this study. Without a rigorous evaluative framework, researchers will continue proposing new LMMs without a reference standard to design against, perpetuating the same methodological weaknesses across generations of models. No LMM will achieve benchmark status not for lack of effort, but because the structural reasons for failure remain undiagnosed. Logistics practitioners, meanwhile, will continue without a credible instrument to assess lean capability consistently across facilities, suppliers, and distribution partners. The study addresses this research gap through four research questions.

RQ1: 

What criteria make a maturity model fit-for-purpose as a measurement instrument?

RQ2: 

How do existing lean maturity models perform against these criteria?

RQ3: 

What critical methodological weaknesses emerge from this evaluation?

RQ4: 

How can future lean maturity models be improved when developing LMMs for logistics and supply chain industries?

To address these research questions, this study first systematically evaluates 27 LMMs against 17 rigor-based criteria derived from measurement theory and maturity model research. In doing so, it introduces the OVRGP framework as an integrated evaluative architecture, representing the first theoretical consolidation of these criteria into a single evaluation framework specifically designed for LMM development and assessment. Second, it derives targeted methodological recommendations to guide future LMM development for logistics and supply chain contexts, addressing the absence of rigorous, logistics-specific lean maturity frameworks identified in the literature. This study adopts a structured review and comparative evaluation design aligned with PRISMA guidelines. RQ1 is addressed through the synthesis of formative measurement theory and maturity model design literature, culminating in the derivation of 17 evaluation criteria grouped under the OVRGP framework. RQ2 and RQ3 are addressed through the systematic comparative evaluation of 27 eligible LMMs against these criteria, using a structured procedure documented in Section 3.1 through Section 3.6. RQ4 is addressed through the analytical synthesis of identified weaknesses and the formulation of targeted improvement recommendations. The first three research questions are addressed at the cross-industry level, as methodological rigor requirements for measurement instruments are independent of industry context. RQ4 narrows to logistics and supply chain contexts to ensure the derived recommendations deliver focused and actionable guidance for a domain where lean maturity assessment remains critically underdeveloped. The study adopts a pragmatic research philosophy, combining qualitative theoretical synthesis with structured empirical evaluation to address a practical and unresolved problem in the lean maturity literature. The resulting framework and recommendations are designed to be immediately useful for researchers, practitioners, and policymakers, while simultaneously establishing the methodological foundations that future studies will require to empirically validate lean maturity models through more positivist approaches.

2. Literature Review

Since the mid-1990s, research has proposed diverse approaches to measure lean [22,23]. Some studies operationalize lean through bundles of practices (e.g., pull systems, 5S), others infer leanness from performance outcomes (e.g., lead time, inventory, set-up time, cost), and a third stream combines enablers and outcomes within composite indices. Systematic reviews document a large and fragmented landscape of these approaches [12,13,14]. A widely used distinction classifies them into: (i) outcome-based assessments, which evaluate improvements achieved after lean implementation (e.g., reductions in lead time, inventories), and (ii) implementation/capability-based assessments, which gauge the degree to which lean principles and practices are embedded [13]. Outcome-based approaches, while useful for reporting purposes, cannot distinguish whether improvements resulted from lean integration or other operational factors. This limitation justifies the existence and continued development of capability-based lean maturity models. However, many such instruments under-specify lean as a multidimensional capability and therefore struggle to capture the systemic integration required for lean to translate into stable, repeatable performance [5,24,25]. Rather than critically evaluating individual models, which is addressed comprehensively through the research questions in Section 4 and Section 5, this literature review establishes the appropriate theoretical measurement model for lean maturity, derives the criteria an LMM should satisfy as a rigorous measurement instrument, critically discusses persistent methodological weaknesses in maturity model development, and identifies the specific gap in logistics and supply chain contexts.

2.1. Specification of the Measurement Model

We conceptualize lean maturity as an abstract latent construct assessed through observable indicators reflecting the extent of lean principle adoption and routinization. Following measurement theory, we argue lean maturity should be modeled formatively: indicators cause the maturity construct rather than reflect it [26,27]. This logic matches lean transformation practice: improvements in selected indicators (e.g., pull replenishment discipline, standardized problem solving, visual performance management, supplier coordination routines) can increase maturity even when other areas remain less developed. Conversely, removing essential indicators can dilute the meaning of “maturity.” Despite its implications for validity testing and interpretation, explicit specification of formative versus reflective logic remains uncommon in LMM studies (with few exceptions such as Leyer and [28]).

2.2. What Criteria Should Maturity Models Satisfy as Measurement Instruments?

This section explores different theoretical criteria MMs should meet to satisfy their purpose as measurement instruments. The fundamental assumption behind MMs is that performance can be improved when the included indicators are improved [29]. This implies the need to establish a clear cause-and-effect relationship between the indicators of an MM and defined outcomes, specifying the business benefits expected when indicators are improved [30,31,32]. Without clarity on desirability, developing an MM becomes meaningless. Among the LMMs, Karlsson and Åhlström [4] and Abdi [33] clearly explain how assessment leads to business improvements, whereas many studies merely state academic relevance and fail to specify business desirability (e.g., [34,35,36]).

2.2.1. Opportunity—Desirability, Scope, Target Group, and Outcome Measures

Lean maturity, as argued earlier, is an abstract latent construct that should be modeled formatively. Given the direction of causality from indicators to the construct, internal consistency is inappropriate for assessing suitability [37]. Instead, validity requires examining how well the construct relates to other variables [37] (p. 333). Diamantopoulos and Winklhofer [26] recommend using an external criterion variable to assess the external validity of formative models. Accordingly, LMMs should specify clear outcome measures derived from their stated business desirability. Some studies comply (e.g., [36,38]), whereas many fail to specify outcome measures (e.g., [33,39,40,41]). A clear definition of the capability domain is critical, as it establishes boundaries for components and indicators [30,42]. Lean has often been interpreted inconsistently, and many LMMs omit explicit scope definition (e.g., [24,43,44]), though some define it clearly (e.g., [4,45,46,47]). Given the complexity of capability domains, the target group must be clearly defined [42]. Defining the target group clarifies the source of information and rater qualifications [30,42]. Some LMMs clearly define target groups (e.g., [25,48]), whereas others do not (e.g., [47,49,50]). When opportunity criteria are absent, the model lacks a defined purpose and the foundation for all subsequent validity and reliability testing collapses.

2.2.2. Validity—Content Specification, Indicator Specification, Construct Validity, and Structural Integrity

Diamantopoulos and Winklhofer [26] outline four criteria for formative validity. First, content specification requires comprehensive coverage within defined domain boundaries. Exclusion of components results in incomplete representation [26,37,51]. Domain components should therefore be clearly specified and justified ([42,52]). Some studies comply (e.g., [22,45]), while others fail to justify inclusion (e.g., [44,48,49]). Second, indicator specification requires coverage of the full breadth of content and theoretically justified linkage to outcome measures [26,27,52,53]. Literature synthesis and expert input support this criterion. Some LMMs justify indicator inclusion (e.g., [35,45]), whereas many do not (e.g., [34,49,50,54]). Third, indicator collinearity may create instability. While variance inflation factor (VIF) is often used to eliminate highly collinear indicators [55], Bollen and Lennox [27] argue that purely statistical elimination may reduce construct meaning. Organizational constructs inevitably involve interdependencies [56]. Accordingly, we do not propose collinearity elimination as a criterion. Fourth, external validity requires meaningful theoretical linkage between indicators and the construct [26,37]. Use of external criterion variables within a MIMIC framework is recommended [57,58]. However, construct validity requires more than specification: upward movement in maturity levels must result in incremental improvements in defined outcomes. Theoretical justification may rely on literature and expert synthesis [42], but empirical construct validity requires longitudinal evidence [59,60]. Many LMMs lack such testing (e.g., [33,34,40,61]). Some claim validation without evidence (e.g., [4,48]). Others compare maturity levels cross-sectionally (e.g., [36,62]), but longitudinal statistical proof remains absent. Structural integrity further contributes to validity. Domain components and indicators should be mutually exclusive and collectively exhaustive (MECE) [42,63]. Overlaps are common (e.g., [34,41,49]). Testing for MECE through theoretical review or expert opinion is therefore essential. When validity criteria are unmet, the maturity scores produced by an LMM cannot be trusted to reflect actual lean capability, making cross-organizational comparison meaningless.

2.2.3. Reliability—Observability, Responsiveness, Language Adequacy, Administrative Procedure, and Repeatability and Reproducibility

Five criteria are critical for reliability: observability, responsiveness, language adequacy, clarity of administrative procedure, and repeatability and reproducibility (R&R). Indicators should be observable to minimize interpretation bias and improve inter-rater reliability [64,65,66]. Some studies use observable metrics (e.g., [4,67]), while others provide insufficient evidence (e.g., [34,36]). Responsiveness refers to the ability to detect change over time [52]. Clear stage definitions enhance responsiveness [31,63,68]. Most LMMs rely on Likert scales, limiting responsiveness, though exceptions exist (e.g., [22,44,47,48,62,69]). Language adequacy strengthens reliability and should be pretested [70,71]. Few studies report such testing (e.g., [24,67]). Administrative procedures should also be clearly specified [63]. Only a minority provide detailed guidance (e.g., [22,25,34,48]). Collectively, these criteria contribute to R&R. Formal R&R testing using Gage R&R [72] or Kappa statistics [73,74] is recommended, yet rarely implemented. Kumar et al. [50] conduct sensitivity analysis, but comprehensive R&R testing is largely absent. When reliability criteria are not met, different raters applying the same model to the same organization will produce different scores, making the instrument unusable for consistent capability tracking or inter-organizational benchmarking.

2.2.4. Generalizability—Proximal Similarity and Cross-Context Validation

Generalizability refers to applicability beyond the original context. Developers should specify proximal contexts [75] and provide thick descriptions [76,77]. True generalizability requires successful application in new contexts with demonstrated improvements [42]. While some studies specify context (e.g., [43,45,61,67,69]), none demonstrate validated cross-context improvements. Nightingale and Mize [48] claim broader results without evidence. Without demonstrated generalizability, an LMM remains a single-context instrument whose findings cannot be transferred, compared, or built upon—severely limiting its value as a reference standard for the field.

2.2.5. Process Integrity—Design Justification and Continuous Improvement

Process integrity requires transparent design justification and evidence of continuous improvement [42,78,79]. Some studies reference design principles (e.g., [22,31,47]), but many provide minimal methodological justification (e.g., [35,38,62]). Evidence of iterative refinement remains limited, with few examples of testing and revision (e.g., [4,48,54,80]). When process integrity is absent, the credibility of the model’s design cannot be independently verified—making its theoretical soundness difficult to assess or replicate.

2.3. Persistent Methodological Weaknesses in Maturity Model Development

Although the number of published maturity models has risen sharply since the CMM [81], they have been consistently criticized for lacking theoretical rigor and empirical evidence of achieving desired outcomes [31,32,78,82,83]. Santos-Neto and Costa [32] report that of journal-published maturity models between 1973 and 2018, only 3% were empirically validated, and only 10% demonstrated real-world application. Wendler [79] further notes that maturity models with proven usefulness remain scarce, and that many have not disclosed the methodological approach taken in their design. While a substantial body of systematic reviews exists [32,60,79,84,85] and standard design processes have been proposed [42,78], studies providing comprehensive design principles and evaluation criteria remain rare [31,86].

2.4. Lean Maturity in Logistics and Supply Chain: An Underdeveloped Research Frontier

Lean management has been widely recognized as a critical organizational capability in logistics and supply chain contexts, where sustained performance improvements depend on the systemic integration of processes, information flows, and operational routines across facilities and partners [6,7,8]. Recent logistics research increasingly frames performance as the result of capability configurations and integration mechanisms spanning facilities and partners, particularly under digital transformation, resilience, and sustainability pressures [15,16,17,18]. Despite this strategic importance, the development of rigorous lean maturity assessment instruments for logistics and supply chain contexts remains severely underdeveloped. Soares et al. [19] explicitly confirm that no validated lean maturity model existed in the supply chain management context using rigorous measurement approaches at the time of their study. Ferraro et al. [21] further demonstrate that maturity models in supply chain management and logistics broadly suffer from limited validation and the absence of comprehensive improvement guidelines. A comprehensive review of lean applications in supply chain management between 2012 and 2024 reveals that existing studies are overwhelmingly focused on isolated SCM activities such as warehouse processes, logistics cost reduction, and manufacturing operations, with no study proposing a lean maturity model designed for logistics and supply chain as an integrated organizational capability [20]. This absence of a rigorous, logistics-specific lean maturity framework directly justifies the recommendations in Section 5.5 of this study.

3. Research Methods

This study is theoretically grounded in formative measurement theory as established in Section 2.1 and in the evaluation criteria derived in Section 2.2, which together constitute the theoretical model informing the research design. To operationalize this foundation, the study first identifies and screens qualifying lean maturity models through a systematic PRISMA-guided search, derives a theoretically anchored evaluation framework through a structured two-stage synthesis process, and applies this framework through a rigorous multi-rater scoring procedure to generate the empirical basis for addressing the four research questions.

3.1. Data Sources and Search Criteria

A systematic search was conducted across four electronic databases: Web of Science, Scopus, SpringerLink, and ScienceDirect. While Web of Science and Scopus are broad indexing databases, SpringerLink and ScienceDirect are publisher platforms and may introduce some coverage bias toward their respective publishing ecosystems. This is acknowledged as a limitation. However, given that the lean maturity model literature is concentrated in operations management and production research journals predominantly indexed across these four databases, the risk of missing qualifying models is considered low. Furthermore, the analytical objective of this study, ‘identifying systematic methodological weaknesses across the LMM landscape,’ is robust to marginal variation in sample size, as the structural patterns identified are unlikely to be reversed by the addition of a small number of models from supplementary sources.

The search was performed using the following executable string, applied across title, abstract, and keyword fields:

(“lean maturity model” OR “degree of leanness” OR “measuring lean adaptation” OR “measuring lean integration” OR “measuring lean level”).

The keyword strategy was designed to directly target studies that explicitly conceptualize and measure the degree of lean implementation, which is the defining characteristic of a maturity-based measurement instrument. The five phrases used collectively cover the core terminology through which lean maturity measurement has been operationalized in the academic literature since 1988. The possibility that a small number of qualifying studies using different terminology were missed is acknowledged as a limitation. Given the consistency of findings across all 27 evaluated models, however, the structural patterns identified are considered robust to marginal variation in sample composition.

3.2. Eligibility Criteria

Studies were included if they:

• Were published in peer-reviewed academic journals, ensuring minimum standards of methodological transparency and scholarly scrutiny consistent with PRISMA-based systematic review practice [87].
• Proposed or substantially developed a structured lean measurement model, reflecting the requirement that maturity models possess defined architectural components, including domain decomposition and scaling logic [42,63].
• Assessed the degree of lean implementation or maturity, rather than only performance outcomes, consistent with the implementation/capability-based assessment category identified in the lean measurement literature [13].
• Contained sufficient methodological detail to evaluate model design and structure, as evaluation against the OVRGP criteria requires minimum reportable evidence of design decisions, indicator specification, and validation procedures [31,63].

Studies were excluded if they:

• Were conference papers, books, book chapters, or non-peer-reviewed publications, as these formats do not provide the level of methodological reporting required for rigorous comparative evaluation [87].
• Measured lean solely through outcome indicators (e.g., cost reduction, lead-time improvement) without assessing implementation maturity, as outcome-only measures do not capture the embedded capability configurations that distinguish maturity-based assessment from performance reporting [13,24].
• Reused existing lean measurement models without significant methodological modification, as inclusion of derivative instruments without independent design contribution would inflate model count without adding evaluative variance [42].

Because terminology varies (e.g., leanness, lean adaptation level, lean integration), models were not excluded solely based on naming conventions.

3.3. Screening and Selection Process

The study selection followed four sequential stages consistent with PRISMA logic [87].

• Identification: Records retrieved from the four databases were compiled, and duplicates were removed.
• Screening: Titles and abstracts were screened to remove clearly irrelevant studies.
• Eligibility: Full-text articles were assessed against inclusion and exclusion criteria.
• Final inclusion: Studies meeting all criteria were retained for detailed evaluation.

An additional structural screening step was conducted to ensure conceptual alignment with the maturity model architecture. To qualify as a lean maturity model (LMM), the study had to demonstrate:

• A first-level construct decomposition (domain components), as maturity models require explicit domain-level structuring to enable meaningful capability differentiation and evaluation [42,63].
• Clearly specified indicators under each domain component, as indicator specification is a minimum requirement for formative measurement instruments to ensure content coverage and construct interpretability [26,63].
• A scaling mechanism reflecting the degree of implementation or maturity progression, as the presence of an explicit scaling logic is the defining architectural feature that distinguishes maturity models from simple assessment checklists [31,63].

Following this process, 27 lean maturity models were retained for analysis.

3.4. Development of the Evaluation Framework (OVRGP) and Embedded Risk-of-Bias Logic

To address RQ1, criteria were derived through a two-stage process. First, a thematic synthesis was conducted across the measurement theory, formative index construction, organizational capability measurement, and maturity model design literature. Source texts were read iteratively and statements specifying what a rigorous measurement instrument or maturity model should demonstrate were extracted and coded into emerging themes. Second, each resulting criterion was traced back to specific source literature to verify theoretical grounding and eliminate criteria that could not be anchored in established measurement or maturity model design principles. This process yielded 17 criteria, which were then consolidated and grouped into five higher-order dimensions based on their conceptual relatedness, termed the OVRGP framework:

○

Opportunity;

○

Validity;

○

Reliability;

○

Generalizability;

○

Process Integrity.

The framework comprises 17 criteria distributed across these five dimensions. Importantly, rather than treating risk of bias as a separate assessment layer, this study integrates risk-of-bias logic directly into the OVRGP structure. Each dimension captures specific forms of methodological bias that may systematically distort maturity assessment.

• OVRGP Dimensions as Bias Domains

Each dimension addresses a distinct category of potential bias:

• Opportunity (Purpose and Desirability Bias)
  Evaluates whether the model clearly specifies its intended logistics or supply chain value creation purpose.
  Bias risk: Absence of explicit business or operational desirability may result in normative or purely academic framing bias, limiting practical relevance.
• Validity (Conceptual and Construct Bias)
  Assesses content specification, indicator justification, formative measurement logic, and linkage to external outcome variables.
  Bias risk: Weak domain boundary definition, incomplete indicators, or lack of empirical construct validation introduces construct validity bias.
• Reliability (Measurement and Administration Bias)
  Examines observability, responsiveness, language adequacy, administrative clarity, and repeatability/reproducibility (R&R).
  Bias risk: Ambiguous indicators, inconsistent rating procedures, or absence of inter-rater testing increase measurement and rater bias.
• Generalizability (Context and Transferability Bias)
  Evaluates contextual boundary definition and evidence supporting applicability beyond the original setting.
  Bias risk: Overgeneralization without contextual specification introduces external validity bias.
• Process Integrity (Design and Reporting Bias)
  Assesses transparency of development methodology, justification of design choices, and evidence of iterative refinement.
  Bias risk: Limited methodological transparency or absence of refinement evidence introduces reporting and design bias.

3.5. Data Extraction, Evaluation, and Embedded Bias Assessment

For each of the 27 included lean maturity models, qualitative evidence was extracted and organized in a structured evaluation matrix (27 models × 17 criteria). Extracted elements included:

○

Model purpose and scope;

○

Domain component definition;

○

Indicator selection and justification;

○

Scaling logic and maturity progression;

○

Validation procedures;

○

Reliability testing;

○

Evidence of contextual specification;

○

Documentation of design and refinement processes.

Each model was rated using a six-point ordinal scale (0–5) for every criterion. Equal weighting was applied across all 17 criteria, reflecting the theoretical position that each criterion represents a necessary and non-substitutable requirement for fit-for-purpose maturity model design. Assigning differential weights would require empirical evidence that certain criteria contribute more to overall model quality than others, evidence that does not currently exist in the maturity model design literature. Equal weighting is therefore the most theoretically neutral and defensible approach given the current state of knowledge [42,63].

Simultaneously, risk-of-bias interpretation was embedded in the scoring logic:

○

Scores of 0–1 typically reflected a high risk of bias within the relevant OVRGP dimension.

○

Scores of 2–3 indicated a moderate risk of bias due to partial fulfillment.

○

Scores of 4–5 indicated low risk of bias supported by clear methodological evidence.

Thus, bias levels were inferred directly from criterion-level compliance rather than assessed separately.

Three raters, each with substantial lean experience as a practitioner, published researcher, and senior academic, evaluated all 27 models independently against the 17 OVRGP criteria using the six-point ordinal scale. Pre-consensus ratings yielded an Intraclass Correlation Coefficient (ICC(2,1)) of 0.836, indicating good inter-rater reliability [88]. Disagreements were concentrated primarily in the Opportunity dimension, where the subjective nature of desirability and outcome specification judgments is most pronounced, with score differences in one rating point in most cases and two points in isolated instances. All disagreements were resolved through structured discussion and re-examination of the original publications until full consensus was achieved. Equal weighting was applied across all 17 criteria, as no empirical basis exists in the maturity model literature for differential criterion weighting [42,63].

3.6. Synthesis and Interpretation

The integrated OVRGP–bias framework enables:

○

Comparative performance assessment across models (RQ2);

○

Identification of systematically high-bias domains (RQ3);

○

Development of targeted methodological recommendations for logistics and supply chain industries (RQ4).

Rather than excluding studies based on bias, the framework interprets bias patterns as structural characteristics of the current LMM landscape. This approach supports constructive advancement of maturity model design in logistics and supply chain research.

4. What Criteria Make a Lean Maturity Model Fit-for-Purpose? The OVRGP Framework

This section addresses RQ1 by introducing the OVRGP framework as the integrated evaluative architecture derived from the theoretical and measurement literature reviewed in Section 2.1 and Section 2.2. The criteria discussed in the literature review are summarized within a conceptual model (Figure 1) and listed in

Table 1. OVRGP list of criteria.

		Criteria Descriptions
1. Opportunity	1.1	Justified desirability 1.1.a: How measurement of maturity in the given domain (e.g., lean management) leads to business benefits should be justified.
	1.2	Defined scope 1.2.a: Scope of the domain should be clearly defined by stating its boundaries. 1.2.b: If similar domains exist, a differentiation should be made (e.g., lean vs. TQM).
	1.3	Defined target group 1.3.a: Target group should be clearly defined according to where data is collected from. 1.3.b: Qualifications of the target group should be stated. 1.3.c: Qualifications of the raters should be stated when self-assessment tools are suggested.
	1.4	Specified outcome measures 1.4.a: Stated business benefits in criteria 1.1 should be converted to measurable outcome variable(s) (e.g., lead time, waste reduction, inventory reduction in lean maturity). These defined outcome variables enable testing the model for construct validity (criteria 2.2 and 2.3).
2. Validity	2.1	Specified content 2.1.a: Domain components that represent the construct (e.g., lean management) should be specified. 2.1.b: Content within each domain component should be clearly defined. 2.1.c: Inclusion of each domain component should be justified based on its potential positive contribution to defined outcome measures in criteria 1.4.
	2.2	Specified indicators 2.2.a: The indicators specified should fully represent the specified domain components specified in criteria 2.1. 2.2.b: Inclusion of the specified indicators should be justified for their potential contribution to improving the outcome measures specified in criterion 1.4.
	2.3	Tested for construct validity 2.3.a: Theoretical construct validity should be achieved by defining the maturity stage definitions in a clear path of meaningful hierarchical progression that leads to improvement in defined outcome variables in criterion 1.4. 2.3.b: Empirical construct validity should be proven with empirical data, that progressing along the stages (maturing) is positively correlated to improvements in the defined outcome variables in criterion 1.4.
	2.4	Tested for structural integrity 2.4.a: All the specified domain components and indicators within them should be tested for mutual exclusiveness and collective exhaustiveness (MECE).
3. Reliability	3.1	Tested for observability 3.1.a: Specified indicators should be defined with observable artifacts.
	3.2	Tested for responsiveness 3.2.a: The stage definitions should be precise in providing the rater with the ability to discriminate between the maturity levels.
	3.3	Tested for language adequacy 3.3.a: The language used in indicators and stage definitions should be understandable to the raters.
	3.4	Defined administrative procedure 3.4.a: User guidelines/administrative procedure on how the maturity model is to be used should be clearly outlined.
	3.5	Tested for repeatability and reproducibility (R&R) 3.5.a: The MM should be tested for R&R before full deployment. Gage R&R techniques used in the six sigma method or Kappa values are recommended.
4. Generalizability	4.1	Theoretical proximal similarity defined 4.1.a: Proximal contexts should be stated where the MM is applicable in improving the specified outcome variables.
	4.2	Empirical proximal similarity proven 4.2.a Construct validity of the MM should be proven within a proximal context outside of where the construct validity has been proven in the first study.
5. Process integrity	5.1	Design process suitability justified 5.1.a: The design process of the MM used should clearly be stated and how the steps within the design process have ensured meeting of criteria should be justified.
	5.2	Model continuously improved 5.2.a: Evidence should be provided on how the MM has been improved as a result of pre-testing or suggestions for improvement should be stated post the use in the empirical study.

with specific attributes to be fulfilled. We define the combined quality achieved by fulfilling these criteria as the fit-for-purpose of an MM. Broadly, for an MM to be fit-for-purpose, it should satisfy two fundamental attributes. First, it should be capable of delivering the desired outcomes within the selected context, meaning it must statistically demonstrate that improved maturity in the selected capability leads to improvements in defined outcomes. Second, it should be generalizable within defined proximal contexts. For an LMM, this implies that the instrument can statistically capture whether maturing in lean management results in improvements in defined outcomes such as customer satisfaction, lead time, or waste reduction, and that similar relationships can be demonstrated in comparable settings such as similar industries.

As illustrated in Figure 1, the ability of an MM to achieve desired outcomes emerges from the combined fulfillment of opportunity, validity, and reliability. These three dimensions are not independent and form a logical dependency chain. Opportunity ensures a clear need, defined scope, specified target group, and explicit outcome variables. These criteria form the foundation for all others. Without defined outcome variables, cause-and-effect relationships cannot be established when testing construct validity; without a defined scope, accurate content and indicator specification are not possible. Validity ensures that the instrument measures what it intends to measure and that its structure allows meaningful correlation between input and output variables, thereby enabling construct validity. Reliability, in turn, depends on validity. An instrument that is not validly specified cannot be meaningfully tested for repeatability and reproducibility, as consistent measurement of a poorly defined construct produces consistently misleading results. Reliability, therefore, requires a validly specified instrument as its prerequisite, and builds upon it through repeatability and reproducibility, supported by observable indicators, unambiguous language, clearly defined administrative procedures, and a responsive measurement scale with explicit maturity stage definitions. Generalizability requires the specification of proximally similar contexts in which the MM can be applied. True reusability is achieved only when empirical construct validity is demonstrated within these defined contexts. Finally, a scientifically credible MM should clearly justify the design process used and document how lessons learned have been incorporated into the published model.

5. Evaluation Results: LMM Performance, Critical Weaknesses, and Improvement Recommendations

This section addresses RQ2, RQ3, and RQ4 using the ratings derived from the systematic evaluation of the 27 selected lean maturity models as described in Section 3. The analysis proceeds in four stages. It first examines the overall performance of all 27 models against the 17 OVRGP criteria, before isolating the highest performing model to understand what enabled its relative strength and where its limitations remain. Cross-model patterns are then analyzed to establish whether the identified weaknesses are structural and field-wide rather than model-specific. The nine criteria recording the lowest average scores are subsequently identified as critical methodological weaknesses, and the section concludes with targeted recommendations to guide future lean maturity model development in logistics and supply chain contexts.

5.1. RQ2—How Do the LMMs Perform Against the OVRGP Criteria?

Figure 2 presents the 27 LMMs ordered chronologically by year of publication. Each cell reports the criterion-level rating (0–5), while the accompanying bar chart displays the overall average score per model across the 17 OVRGP criteria.

• Overall Trends

Despite more than 25 years of LMM development, the findings reveal that methodological rigor remains consistently low across the evaluated models. Even the most recent models do not demonstrate consistently strong methodological fulfillment.

Importantly:

○

No LMM achieved an overall average rating of 3 or above.

○

Performance variability across criteria remains moderate.

○

Standard deviation patterns suggest that most models exhibit systematic weaknesses across multiple dimensions, rather than isolated deficiencies.

This pattern indicates structural limitations within the LMM landscape rather than generational improvement.

5.2. Highest-Performing Model and Its Limitations

Among the evaluated models, Malmbrandt and Åhlström [22] achieved the highest overall average score. Their lean service maturity model demonstrates relatively strong performance within the Opportunity dimension, particularly in:

○

Justifying business desirability;

○

Defining outcome measures;

○

Articulating a coherent scope.

The model development process combined theoretical grounding with iterative practical validation, strengthening its Process Integrity dimension relative to other LMMs. The use of generic maturity stage definitions (adapted from Nightingale and Mize [48]) improved responsiveness compared to models relying solely on Likert-type scales.

Nevertheless, critical weaknesses remain:

○

Construct validity relies primarily on expert opinion rather than empirical demonstration of maturity–outcome relationships.

○

Structural integrity (MECE) is not formally tested.

○

The inclusion of lean enablers, practices, and performance indicators within the same structural layer raises concerns of conceptual overlap and embedded cause-and-effect circularity.

Thus, even the strongest-performing LMM presents a moderate risk of bias under the integrated OVRGP–bias framework.

5.3. Cross-Model Patterns and Systematic Weaknesses

Across the 27 LMMs, several recurring patterns emerge:

1.

Opportunity is partially satisfied in many models, yet outcome variables are frequently underspecified, limiting construct testing.

2.

Validity weaknesses are widespread, particularly regarding:

○

Empirical construct validation;

○

Justification of indicator inclusion;

○

Testing of structural integrity (MECE).

3.

Reliability is inconsistently addressed, with limited evidence of:

○

Observability testing;

○

Inter-rater reliability assessment;

○

Repeatability and reproducibility (R&R) testing.

4.

Generalizability remains the weakest dimension, as very few models demonstrate empirical validation beyond their original application context.

5.

Process Integrity is often underreported, with limited transparency in design logic or iterative refinement.

The relatively narrow standard deviation across criteria indicates that most LMMs perform at similar moderate-to-low levels of methodological rigor. In other words, the field demonstrates consistency in structural limitations, rather than isolated excellence.

In formative measurement, the structural stability of an index depends on the fulfillment of specific foundational requirements—valid construct specification, justified and non-overlapping indicators, empirically tested outcome relationships, and documented reliability across raters and contexts [26,27,37]. When these requirements are systematically absent, as the OVRGP scores demonstrate across the majority of the 27 evaluated models, the resulting index scores are mathematically computed but theoretically compromised. The maturity scores such models produce cannot be trusted to reliably reflect actual lean capability, nor can they support credible comparison across organizations or contexts. It is in this precise sense that the identified weaknesses are structural rather than incidental.

Figure 3 further disaggregates criterion-level performance to identify systematically weak dimensions, providing deeper insight into RQ2 and setting the stage for RQ3.

Figure 3 presents the average rating and average standard deviation for each of the 17 criteria, aggregated by OVRGP category. This disaggregated analysis provides deeper insight into systematic strengths and weaknesses across the LMM landscape.

• Criterion-Level Performance Patterns

Among the five OVRGP dimensions, Opportunity records the highest average score. Most LMMs articulate business desirability, scope, and target audience at least partially. However, the specification of measurable outcome variables remains inconsistent, as only a subset clearly translates intended benefits into operationalized performance metrics. The comparatively stronger performance of Opportunity reflects the conceptual nature of its criteria. Defining scope, target group, and intended purpose does not require empirical testing or statistical validation. This pattern also applies to other criteria grounded primarily in conceptual articulation rather than empirical demonstration, including:

○

Content specification;

○

Indicator specification;

○

Clarity of administrative procedures;

○

Theoretical definition of proximal context;

○

Justification of design logic.

These criteria show comparatively higher average ratings across models.

• Empirical Rigor as a Systematic Weakness

In contrast, criteria that require empirical validation, statistical testing, or structured expert involvement consistently exhibit low average ratings. These include:

○

Empirical construct validity;

○

Structural integrity testing (MECE);

○

Language adequacy testing;

○

Empirical proximal similarity (cross-context validation);

○

Demonstrated continuous improvement of the model.

The consistently low performance in these areas suggests that most LMMs remain at the conceptual stage and do not advance to rigorous validation. This raises a broader methodological concern: the field appears largely driven by literature synthesis and author judgment, with limited empirical verification of maturity–outcome relationships. From the perspective of the integrated OVRGP–bias framework, this pattern indicates elevated construct, measurement, and transferability bias across the majority of models.

• Variability Across Criteria

Standard deviation values across criteria are moderate and stable, indicating that weaknesses are broadly distributed rather than confined to a few models. An exception is the “specification of outcome measures,” which shows higher variability, reflecting a polarized pattern in which some authors operationalize outcomes clearly while others omit them entirely. As expected, the lowest standard deviation appears in “empirical proximal similarity,” where nearly all models received minimal or zero ratings, revealing a near-universal absence of cross-context empirical validation. Overall, the findings from Figure 3 suggest that methodological limitations of LMMs are structural rather than incidental: conceptual articulation is common, whereas empirical substantiation is rare. Strengthening empirical validation and cross-context testing, therefore, represents a critical pathway for improving the fit-for-purpose of future lean maturity models in logistics and supply chain research.

5.4. RQ3—What Are the Critical Methodological Weaknesses?

In addressing RQ3, we identify nine criteria within the OVRGP framework that constitute critical methodological weaknesses. Figure 4 presents all 17 criteria ranked in descending order according to their average rating across the 27 evaluated LMMs.

The nine lowest-performing criteria score below an average of 2 on the six-point scale, dividing the 17 criteria into two nearly equal groups. Notably, all eight higher-ranked criteria share a common feature: they do not require empirical validation, statistical testing, or structured expert assessment. This reinforces the earlier observation that LMM literature has largely emphasized conceptual articulation over empirical substantiation. Although these comparatively stronger criteria are not classified as critical weaknesses, their performance remains modest, with most averages between 2 and 3, indicating only partial fulfillment. The only exception is “justified desirability,” which scores slightly above 3. Even in the strongest conceptual areas, methodological rigor remains moderate rather than robust.

The nine critical weaknesses predominantly correspond to criteria requiring:

○

Empirical construct validation.

○

Structural integrity testing (MECE).

○

Reliability testing (e.g., R&R, language adequacy.

○

Empirical cross-context validation.

○

Demonstrated iterative refinement.

This clustering indicates that the most demanding aspects of measurement design, those requiring data collection, statistical analysis, or systematic validation, are consistently underdeveloped. At least one critical weakness appears within each of the five OVRGP dimensions, showing that limitations of existing LMMs are structurally distributed across purpose articulation, validity, reliability, generalizability, and process integrity. From a measurement theory perspective, this is concerning. A maturity model intended as a scientific instrument must demonstrate credible construct validity, reliability, and transferability. Weak performance across these domains reduces the overall fit-for-purpose of current LMMs and limits their usefulness for evidence-based logistics and supply chain decision-making. The following subsections examine these nine critical weaknesses in detail and propose targeted methodological improvements, addressing both RQ3 and RQ4.

5.5. RQ4—How Can Future Lean Maturity Models Be Improved When Developing LMMs for Logistics and Supply Chain Industries?

Figure 5 presents the 27 LMMs ranked in descending order according to their overall average scores. When performance is examined across the nine identified critical weaknesses, a consistent pattern emerges: most models score zero or near zero, with only isolated exceptions achieving ratings above 4 or 5. Among the weakest-performing criteria, two are particularly critical for logistics and supply chain applications: repeatability and reproducibility (R&R) and empirical proximal similarity.

• Repeatability and Reproducibility (R&R)

The absence of explicit R&R testing raises substantial concerns regarding the use of LMMs as measurement instruments in logistics and supply chain environments, where assessments often span multiple facilities, functional units, and inter-organizational interfaces. A model that cannot demonstrate consistent results across raters, locations, or time lacks the stability required for network-level capability evaluation. A rating of zero does not imply inherent unreliability, but rather the absence of documented evidence of systematic testing (e.g., inter-rater reliability, intra-rater consistency, Kappa statistics, or Gage R&R procedures). In distributed logistics systems, where maturity may be assessed across warehouses, distribution centers, transport nodes, or supplier interfaces, the omission of documented R&R testing increases measurement bias risk within the Reliability dimension of the OVRGP framework. Without formal evidence of rating stability, conclusions regarding maturity differentiation across supply chain entities remain methodologically vulnerable.

• Empirical Proximal Similarity

The second lowest-performing criterion, empirical proximal similarity, concerns whether construct validity holds beyond the original study context. In logistics and supply chain research, where operational configurations vary across industries, regions, and network structures, demonstrating cross-context validity is essential. While several models describe theoretically applicable settings, very few provide empirical evidence that maturity–outcome relationships are reproducible in comparable logistics or supply chain environments. Merely specifying potential application domains does not establish reusability. To function as transferable instruments, LMMs must show that progression in lean maturity correlates with improved operational or network-level outcomes in at least one additional proximal context (e.g., similar distribution networks, manufacturing supply chains, or service-based logistics systems). Unsurprisingly, empirical proximal similarity scores are uniformly low, particularly given the limited evidence of empirical construct validation even within original contexts.

Although cross-context validation may extend beyond the initial development phase of an LMM, from a scientific measurement perspective, it is indispensable for establishing broader acceptance in logistics and supply chain scholarship. Excluding this criterion would weaken the evaluative rigor of the OVRGP framework and lower the standards required for maturity models intended to assess inter-firm and multi-site lean capability development. Accordingly, empirical proximal similarity is retained as a forward-looking benchmark for methodological maturity in future LMM development within logistics and supply chain research.

In addressing RQ4, we have first examined the consequences for the fit-for-purpose of LMMs in logistics and supply chain applications when these weaknesses remain unaddressed. We then provide methodological recommendations intended to guide future researchers in strengthening the design and validation of lean maturity models.

This dual perspective moves the discussion beyond critique toward constructive advancement of lean maturity assessment in supply chain contexts. From the first perspective, each unresolved weakness undermines one or more OVRGP dimensions. For instance, the absence of empirical construct validation weakens the credible linkage between lean capability development and operational outcomes such as service reliability or flow efficiency (Validity dimension). Lack of R&R testing compromises measurement stability across sites and evaluators (Reliability dimension), while failure to demonstrate empirical proximal similarity limits transferability across comparable supply chain configurations (Generalizability dimension). Collectively, these deficiencies reduce the overall fit-for-purpose of LMMs as scientific instruments in logistics and supply chain systems.

From the second perspective, these weaknesses translate into targeted methodological recommendations (

Table 2. Synthesized recommendations for each critical weakness.

Specified outcome measures

Incremental lean maturity can lead to improvements in outcomes such as waste reduction, business growth, lead-time improvement, and enhanced employee and customer satisfaction. When developing industry-specific LMMs, outcome measures should be customized to reflect sector realities, for example, perfect order rates, customer lead times, and cash-to-cash cycle time in logistics and supply chain industries.

Tested for construct validity

It remains unclear whether lean maturity is conceptually well established within the logistics and supply chain literature. Although multiple measurement approaches exist, current best practice favors generic maturity stage definitions that were largely developed in manufacturing contexts. Further research should aim to define a meaningful maturity path grounded in the progression of organizations advanced in lean implementation within logistics and supply chain environments. Toyota is often cited as the lean benchmark, yet its most instructive lessons for logistics practitioners may lie not in its production system but in its supplier development practices, inbound logistics coordination, and distribution network discipline, areas where lean maturity progression is observable and consequential. A study integrating Toyota’s supply chain journey with lessons from other organizations advanced in logistics lean capability, validated through qualified expert input, is recommended to define maturity stages for each domain component or indicator. Where such evidence is limited, generic stage definitions may serve as a pragmatic alternative. To establish construct validity, logistics and supply chain enterprises should be examined longitudinally across statistically sufficient rounds to test the relationship between incremental maturity and network-level outcome improvements. This requires systematic collection of both indicator data, such as problem-solving rates, visual management usage, lean tool application, and leadership engagement, and logistics-specific outcome data, such as perfect order rates, delivery lead times, inventory turns, and waste reduction across supply chain interfaces. Although surveys may appear convenient, they are weak in addressing observability, responsiveness, and repeatability and reproducibility. Instead, onsite observations, objective evidence collection, and validation through structured interviews across multiple supply chain sites are recommended to strengthen measurement rigor.

Tested for responsiveness

As discussed earlier, the concept of maturing in lean management has not been fully understood, and this is particularly true in logistics and supply chain contexts where lean maturity research remains nascent. In such a context, it would be difficult to establish stage definitions with high responsiveness. However, developers of LMMs can pre-test their proposed stage definitions before using them in the study. A sample of lean maturity scenarios representing different indicators, such as visual management, use of lean tools, problem-solving routines, and supply chain coordination practices, can be used to test stage definitions using a sample of qualified raters drawn from logistics and supply chain functions.

Tested for observability

Many existing LMMs define indicators at a conceptual level without specifying observable artifacts, for example, smooth information flow, team-based decision-making, and personnel interchangeability. Such abstraction makes a consistent rating difficult. Instead, indicators should be anchored in observable artifacts that enable reliable assessment. In logistics and supply chain contexts, these may include visible performance management boards at warehouse and distribution center levels, daily operational huddles across shift teams, documented root cause validation records, synchronized replenishment signals with suppliers, transparent leadership coaching schedules, cross-site lessons learned documentation, structured employee-led problem-solving practices, and supplier delivery performance review records.

Tested for structural integrity

LMMs employ diverse domain components, reflecting a lack of consensus on how lean management should be structurally decomposed. Many models use tools and practices such as JIT and Kanban as domain components, although these are not mutually exclusive and are often hierarchically embedded. Such structuring compromises MECE and limits scientific usability. Domain components should instead be grounded in lean principles, which are inherently more mutually exclusive and collectively exhaustive. Widely recognized alternatives include the four-P model [6], the five lean values [2], and Liker’s 14 principles [89]. In logistics and supply chain contexts, these principle-based frameworks can be meaningfully translated into domain components such as inbound flow management, inventory discipline, outbound fulfillment, supplier integration, information flow, and continuous improvement culture, each of which represents a distinct and non-overlapping capability area. Once expanded into indicators, MECE should be tested through a qualified expert assessment by asking whether the indicators fully represent the domain component and whether they are mutually exclusive. Indicator refinement should continue until an acceptable expert consensus is achieved.

Tested for language adequacy

As a first step, we recommend converting any lean jargon to the normal business language used in the industry context where the LMM is developed. For example, ‘the distribution center operates with heijunka-based inbound scheduling’ can be understood differently even among qualified lean experts. This can be written as ‘the distribution center receives and processes inbound shipments according to a levelled daily schedule that matches downstream demand’. Once written, these indicator statements should be tested for language adequacy by using a sample of raters before the actual investigation.

Model continuously improved

We recommend that the studies clearly state the steps taken at each phase of the design process to improve the overall fit-for-purpose. For example, showing what was refined after pre-testing the instrument for language adequacy, responsiveness, RandR, etc., and listing the weaknesses identified from the empirical study.

Tested for repeatability and reproducibility (RandR)

The recommendations provided under observability, responsiveness, and language adequacy directly contribute to ensuring increased RandR. Once the indicators are tested for these three criteria, we recommend an RandR test to verify the final measure of reliability. A set of sample scenarios drawn from realistic logistics and supply chain settings can be used among a sample group of qualified raters to pre-test for RandR. As mentioned in Table 1, Kappa values or gauge RandR tests used in the six sigma method can be used. Further, we recommend that the actual investigation be conducted with multiple raters across different supply chain functions and facility types, as this would allow testing of inter-rater reliability on the actual collected data and reveal any systematic rating differences across warehousing, transport, and supplier management contexts.

Empirical proximal similarity proven

We recommend that this criterion be achieved via a follow-up study within a specified proximal context by establishing construct validity. In logistics and supply chain research, the original study should specify the operational context for which the LMM is developed, for example, a lean maturity model for third-party logistics providers, and prove its construct validity within a selected sample of comparable organizations. To prove generalizability, a follow-up study could be conducted in a different but proximal context, for example, a similar distribution network configuration, the same logistics sector in a different geographical setting, or a comparable supply chain tier, such as tier-one supplier networks. We agree that this criterion is a relatively high expectation from an academic instrument. However, in logistics and supply chain environments where operational configurations vary considerably across industries, regions, and network structures, demonstrated cross-context validity is what distinguishes a transferable measurement instrument from a single-use diagnostic tool. Such an achievement will increase not only the academic credibility of the instrument but also its acceptance among logistics practitioners and consulting professionals seeking reliable benchmarking tools.

The analysis of the nine critical weaknesses demonstrates that current LMMs are conceptually articulated but empirically underdeveloped. While many models define scope and structure, few provide robust evidence of construct validity, reliability, or cross-context applicability. As shown in Table 2, strengthening lean maturity assessment requires systematic attention to outcome specification, longitudinal validation, structural integrity, and formal reliability testing. Addressing these gaps is essential for transforming LMMs from descriptive frameworks into scientifically credible instruments capable of supporting evidence-based logistics and supply chain decision-making.

). We advocate systematic strengthening of model development practices, including clearer specification of supply chain-relevant outcome variables, formal MECE testing, longitudinal empirical validation, documented reliability assessment across raters and sites, cross-context replication, and transparent reporting of iterative refinement.

6. Discussion

6.1. Theoretical Implications

The OVRGP framework addresses this gap by consolidating 17 rigor-based criteria derived from formative measurement theory and maturity model design literature into a single integrated evaluative architecture. The consistent pattern of low empirical substantiation across 27 models spanning more than 25 years of development suggests that the identified weaknesses are structural and field-wide rather than generational or model-specific. This study, therefore, provides the theoretical foundation necessary for transforming LMMs from descriptive frameworks into validated measurement instruments.

6.2. Implications for Managers, Practitioners, and Policymakers

The OVRGP framework and its associated recommendations serve as a practical reference for organizations, consulting practitioners, and policymakers seeking to develop or select rigorous lean maturity instruments for logistics and supply chain transformation. Practitioners can use the framework diagnostically to assess the methodological credibility of existing LMMs before adopting them for capability assessment. Organizations commissioning LMM development can use the 17 criteria and nine recommendations in Table 2 as a design standard to ensure the resulting instrument meets minimum scientific requirements. Policymakers can use the OVRGP framework as a reference architecture for developing standards and certification programs for lean capability assessment across critical supply chain processes, supporting more consistent and credible benchmarking at the industry and network level.

6.3. Limitations and Future Research Directions

This study acknowledges several limitations. The keyword search strategy and database coverage, while systematically applied, may have missed a small number of qualifying studies using different terminology, though the consistency of findings across 27 models suggests the structural patterns identified are robust to marginal variation in sample composition. Disagreements among raters were concentrated in the Opportunity dimension, reflecting the inherently subjective nature of desirability and outcome specification judgments. The recommendations derived from this study are theoretically grounded but not yet empirically validated. Their practical effectiveness will only be demonstrated when a purpose-built logistics-specific LMM is developed and tested using this theoretical foundation, representing a distinct and necessary avenue for future research. The OVRGP framework itself also awaits empirical validation against a demonstrably rigorous LMM to confirm that fulfilling all 17 criteria produces a superior instrument in practice. Future research should focus on developing and empirically validating a logistics-specific lean maturity model grounded in the OVRGP criteria, testing cross-context generalizability across comparable supply chain configurations, and longitudinally validating the relationship between lean maturity progression and network-level performance outcomes.

7. Conclusions

This study systematically evaluated 27 lean maturity models against 17 rigor-based criteria grouped under the OVRGP framework, revealing nine critical methodological weaknesses distributed across construct validation, reliability testing, structural integrity, and empirical generalizability. The consistent pattern that emerges across more than 25 years of LMM development is unambiguous: the field has prioritized framework construction over measurement validation, producing instruments that are conceptually articulated but empirically underdeveloped. The OVRGP framework introduced in this study provides the first theoretically consolidated reference architecture for evaluating and developing rigorous lean maturity models, with targeted recommendations specifically designed for logistics and supply chain contexts. Strengthening methodological rigor is not an optional refinement. It is a prerequisite for lean maturity models to achieve scientific credibility and practical utility as instruments of organizational capability assessment.