Information-Enabled Marketing Efficiency and Financial Performance in Centralized Finance (CeFi)—An International Study

Abstract

This study examines the statistical associations between commercialization-related cost structures and financial outcomes on revenue growth, profitability, and scale within a centralized financial system. We estimate four OLS models (M1–M4) using aggregated annual data from 2020 to 2025 and enhance our analysis with a fuzzy cognitive map (FCM) scenario assessment. The findings demonstrate that revenue growth correlates positively with both SG&A growth and commercialization efficiency (revenue per unit of SG&A); however, SG&A intensity exhibits a negative relationship with net margins. Logarithmic estimations indicate a robust co-scaling between operational expenses and revenues, implying growth driven by capacity rather than operating leverage. Lagged analysis also reveals an intertemporal trade-off, wherein phases of accelerated SG&A expansion are succeeded by diminished subsequent growth. The findings underscore the necessity of differentiating between commercialization intensity and efficiency, and advise against viewing SG&A growth as a consistent alignment of financial performance.

1. Introduction

Centralized finance (CeFi) intermediaries, including custodial exchanges, brokerages and centrally governed digital financial service providers, compete in markets where growth depends on user acquisition, trust formation and sustained platform engagement [1]. Unlike decentralized finance (DeFi), CeFi typically retains control over custody, compliance and customer interfaces, so its cost structure and performance drivers resemble those of regulated intermediaries more than purely protocol-based systems [2]. Competitive interactions and spillovers between DeFi and CeFi lending markets further suggest that CeFi growth unfolds under cross-ecosystem pressure rather than in isolation [3]. Accordingly, this study treats CeFi as institutionally governed financial intermediation delivered through platformized customer interfaces, which enables analysis using auditable accounting disclosures.

In this setting, information systems (ISs) are not peripheral infrastructure. They shape how commercialization is executed, monitored and scaled. Digital transformation (DT) is commonly conceptualized as an organizational change process enabled by combinations of information, computing, communication and connectivity technologies [4,5]. Prior evidence indicates that IS capabilities contribute to performance, largely through mediating improvements in decision-making quality and business process execution [6]. However, the magnitude, timing and measurement of these effects are context-dependent, particularly in financial services where onboarding friction, risk controls and compliance workflows directly affect conversion and retention [7,8].

A parallel debate in marketing and strategy concerns accountability. Firms are pressed to demonstrate how customer-facing investments translate into financial outcomes even when attribution remains difficult despite abundant digital trace data. Marketing performance assessment (MPA) research emphasizes that decision usefulness depends on the design and governance of measurement systems, especially metric selection, interpretation and integration into planning and control [9], rather than the volume of available indicators alone [10]. For CeFi, this challenge is amplified because commercialization spans advertising and promotion, and also IS-mediated onboarding, customer support and assurance-related experience investments that influence trust and continued use.

Empirical analysis is complicated by the lack of clean marketing disclosures in financial statements. Customer acquisition activities, sales operations, support capacity and overhead frequently co-mingle within Selling, General & Administrative (SG&A) expenses, making SG&A a pragmatic but conceptually noisy proxy for commercialization resourcing. Comparative evidence shows that SG&A-based marketing proxies can yield inferences that diverge from models based on disclosed marketing expenditures, underscoring construct validity risks [11]. In addition, SG&A may exhibit asymmetric adjustment (cost stickiness), so SG&A growth can reflect managerial commitment and adjustment frictions rather than purely contemporaneous commercialization choices [12].

Despite extensive research on marketing accountability and digital transformation, empirical evidence linking commercialization-related expenditures to financial performance in centralized financial intermediation remains limited. In CeFi, commercialization, compliance infrastructure, customer support and governance execution are not reported as distinct line items but are embedded within SG&A and operating expense aggregates. This aggregation constrains construct clarity: SG&A cannot be interpreted as pure marketing expenditure, yet it remains the only auditable, cross-firm comparable signal of commercialization-related resource commitment.

Existing studies either examine marketing expenditure in non-financial sectors or analyze digital transformation effects without isolating commercialization dynamics within regulated digital financial intermediaries. Consequently, the literature does not yet clarify whether accounting-based commercialization indicators, specifically SG&A-derived efficiency, intensity and expansion dynamics, provide meaningful explanatory signals for growth and profitability in centralized digital financial intermediaries. This measurement constraint constitutes the central empirical gap addressed in this study.

In this paper, the term “information-enabled” is used in two complementary senses. First, the analysis relies on auditable accounting information (SG&A and operating cost aggregates) as the only consistently comparable signals available for accountability across firms and years. These indicators are treated as outputs of measurement systems rather than as clean observations of marketing spend. Second, the study uses these signals to construct an explicit decision-support representation based on fuzzy cognitive maps (FCMs) that makes assumed linkages inspectable and supports joint what-if reasoning. Accordingly, the study does not claim to measure IS capability maturity directly. Instead, it examines how the bundled commercialization and operations posture captured in financial disclosures aligns with growth, margin and scale outcomes.

Despite substantial research on digital transformation and marketing accountability, three specific gaps remain insufficiently addressed in the CeFi context. First, while digital transformation and performance relationships have been examined in banking and financial services settings [8,13], prior studies have rarely investigated commercialization–performance linkages in centralized digital financial intermediaries where marketing, compliance, and governance costs are structurally bundled within SG&A aggregates, creating interpretive constraints not present in sectors with disaggregated marketing disclosures [11,12]. Second, although marketing productivity research differentiates input–output conversion efficiency from expenditure levels [9,10], existing work does not explicitly examine how commercialization productivity (conversion efficiency) and commercialization intensity (cost density) operate under accounting aggregation constraints in regulated financial systems characterized by classification heterogeneity and cost behavior frictions [11,14]. Third, while regression-based evidence on marketing–performance relationships is well developed [9,10], the translation of empirically observed associations into structured scenario-based decision tools remains limited, particularly in financial service contexts with interdependent cost structures, despite growing interest in FCM-based decision-support modeling for complex systems [15,16].

Accordingly, this study aims to examine whether SG&A-derived commercialization productivity (revenue/SG&A), commercialization intensity (SG&A/revenue) and SG&A expansion dynamics are systematically associated with revenue growth, profitability and revenue scale in CeFi. By distinguishing commercialization efficiency, intensity and expansion dynamics under accounting aggregation constraints, the study clarifies how SG&A-based indicators should be interpreted in accountability systems within centralized digital financial intermediaries. This also provides a structured way to link commercialization signals embedded in financial disclosures with decision-support reasoning about growth, profitability and operational capacity. Beyond estimating linear associations, the study translates the empirically observed relationships into a structured “what-if” scenario framework using fuzzy cognitive map (FCM), enabling system-level exploration of how simultaneous shifts in commercialization momentum and operating capacity may affect financial outcomes [15,17].

This study advances theory and practice in three specific ways. First, it contributes to the marketing accountability and measurement literature by explicitly disentangling commercialization efficiency (revenue/SG&A) from commercialization intensity (SG&A/revenue) under accounting aggregation constraints. While prior research acknowledges that SG&A is an imperfect proxy for marketing expenditure and may introduce construct validity concerns [11,12], few studies examine how distinct SG&A-derived constructs operate simultaneously within regulated, centralized financial systems where marketing, compliance, governance, and service delivery costs are structurally bundled. By separating productivity-type conversion measures from cost burden measures and expansion dynamics, the study refines how accounting-based commercialization indicators should be interpreted in accountability systems [10].

Second, the study extends digital transformation and financial intermediation research by positioning CeFi growth within a capacity co-scaling framework rather than an operating leverage framework. The near-unit elasticity between operating expenses and revenue suggests proportional scaling of compliance, service, and infrastructure capacity rather than costless digital expansion. This finding contributes to debates on performance measurement and digital scaling in regulated environments, where measurement design and structural cost behavior matter for interpretability [10,12].

Third, the study contributes methodologically by translating regression-based accountability evidence into a structured decision-support representation through a data-informed fuzzy cognitive map (FCM). While regression models establish conditional associations, managerial environments require system-level reasoning under joint shocks. FCM-based modeling supports inspectable and bounded scenario exploration in interdependent systems [17], thereby operationalizing empirically observed sign structures into a decision-support logic consistent with information systems accountability research [10]. Rather than serving as a forecasting device, the FCM layer provides structured sensitivity analysis aligned with calls for actionable performance measurement architectures [10,17]. The remainder of the paper synthesizes the relevant literature and develops hypotheses, describes the research design and variable construction, presents empirical results and discusses the implications for CeFi research and practice.

2. Literature Review

2.1. IS-Enabled Trust, Governance and Commercialization in CeFi

CeFi commercialization is inseparable from IS-enabled governance and assurance because acquisition and retention depend not only on access and pricing, but also on how custody, compliance and service delivery are executed through digital processes. The key implication of centralized custody and governance is that platform design choices create both constraints and opportunities for commercialization [2]. In parallel, documented spillovers and substitution dynamics between DeFi and CeFi lending markets suggest that CeFi growth and engagement should be interpreted against an evolving market architecture rather than as the product of outward-facing promotion alone [3]. From an IS perspective, trust is therefore best treated as an operationally produced outcome, shaped by system reliability, disclosure practices, incident response and customer support, alongside the credibility of assurance mechanisms—rather than as a purely reputational asset [1,18].

Digital transformation (DT) scholarship further underscores that performance-relevant change typically arises through the reconfiguration of routines, decision rights and process execution enabled by integrated information, computing, communication and connectivity technologies [4,5]. In CeFi, commercialization is executed through IS-intensive end-to-end customer-journey systems, onboarding and verification pipelines, attribution and experimentation, fraud and risk analytics, customer service orchestration and compliance workflow automation. Consequently, growth, engagement and monetization are co-produced by information architecture and process governance, not solely by front-end communication.

Empirical IS capability research supports a mechanism-based expectation: IS capabilities rarely improve performance directly but do so primarily through mediating improvements in decision-making performance and business process performance [6]. This mechanism is salient in CeFi because customer experience is jointly determined by speed (reduced friction), assurance (reliability and safety) and regulatory execution (e.g., KYC/AML), all of which are process-intensive and data-dependent. Accordingly, treating commercialization narrowly as communication activity risks obscuring the operational channels through which CeFi platforms convert market-facing effort into scalable revenue.

Marketing- and capability-oriented research complements this view by showing that firms often pursue digital commercialization goals without mature digital marketing capabilities and that capability gaps can constrain performance even when digital tools are available [19]. Related evidence suggests that digital marketing capabilities can be value relevant, but their contribution depends on how they are embedded within broader resources and decision systems rather than treated as isolated “front-end” tools [20]. For CeFi, this implies that scalable commercialization likely requires the coordinated development of market-facing capabilities (e.g., targeting and channel orchestration) and assurance capabilities (e.g., uptime, transaction reliability, and dispute resolution) because users experience trust through repeated system-mediated interactions.

Evidence from financial services also cautions against assuming monotonic benefits from digital transformation. Large-sample studies report positive associations between DT and profitability, while highlighting substantial heterogeneity by institutional setting, measurement design and the dimension of digitalization captured [8,13]. For the present study, the implication is that the commercialization–performance linkage should be treated as an empirical question once both capability benefits and operating costs are considered. This motivates the accountability-oriented perspective developed next, which clarifies how commercialization-relevant resource commitments can be assessed in a way that remains interpretable for managerial decision-making.

Because CeFi institutions operate within shared regulatory constraints and digital infrastructure environments, this study adopts an aggregated sector-level perspective to capture systemic commercialization dynamics rather than firm-level heterogeneity.

2.2. Marketing Performance Assessment, Measurement Systems and Accountability in Digital Contexts

A central challenge in studying commercialization is that it is simultaneously a managerial activity (resource allocation and execution) and a measurement problem (how outcomes are attributed and evaluated). Contemporary marketing scholarship treats accountability less as a question of whether metrics exist and more as a question of how marketing performance assessment (MPA) is designed, governed and used in decision-making. The MPA process perspective highlights that assessment systems can fail even when data are abundant because bottlenecks often involve metric selection, interpretation, incentives and integration into planning and control routines [10]. This is especially pertinent in digital-first settings where trace data can create an illusion of control while leaving attribution, comparability and decision usefulness unresolved.

This concern is formalized in the literature on marketing performance measurement systems (MPMSs), which conceptualizes measurement as an organizational system rather than a reporting layer. Evidence suggests that MPMS effectiveness depends on both measurement design and how the system is used, with marketing capabilities providing an important link between measurement and performance [21]. Moreover, recent work distinguishes diagnostic uses of metrics (monitoring against targets) from interactive uses (strategic dialog, learning and adaptation), showing that performance effects can be nonlinear and contingent on competitive conditions and metric focus [22]. For platform-mediated financial services, this implies that accountability depends on whether metrics support learning and reallocation under uncertainty rather than merely documenting outcomes ex post.

A second strand of research examines the “right metric for the right decision” problem. In marketing-mix decisions, the choice of metric (e.g., short-term financial outcomes versus customer-mindset metrics) is associated with perceived decision performance, and mismatches between metric and decision context can reduce effectiveness even when a metric is technically valid [9]. This matters in commercialization settings where decisions are multi-objective (acquisition, retention, trust-building and experience execution) and where short-horizon metrics can bias attention toward what is measurable rather than what is strategically consequential. Accountability therefore requires explicit alignment among decision horizon, metric controllability and the organizational level at which the decision is made.

Digitalization and automation intensify these tensions. Advanced analytics and AI can expand what is measurable and optimize execution, but they also introduce opacity, feedback effects and governance challenges that complicate interpretability. Conceptual work on AI in marketing emphasizes that automation benefits depend on integration with managerial control, strategic intent and customer relationship objectives rather than stand-alone tool deployment [23]. From an accountability standpoint, increasingly sophisticated optimization can reduce transparency about why performance changes, raising the value of explainable and decision-relevant measurement architectures.

Taken together, these studies motivate the use of interpretable and comparable commercialization indicators when granular marketing spend disclosures are unavailable. MPMS/MPA evidence implies that accountability mechanisms should support learning and resource allocation while remaining robust to attribution noise [10,21]. For the present study, accounting-level indicators are not treated as ideal measures of marketing, but as baseline accountability signals that remain auditable and comparable across firms and years. This framing sets up the next section’s focus on operationalizing commercialization through accounting-based proxies while maintaining strict discipline about what such measures can and cannot identify.

2.3. Operationalizing Commercialization with Accounting Proxies: SG&A, Efficiency vs. Intensity and Construct Validity

Empirical research on commercialization in platform-mediated financial services faces a persistent constraint: comparable disclosures of marketing expenditures are uncommon across firms and years. As a result, researchers often rely on auditable accounting aggregates, most commonly SG&A expenses, to proxy commercialization-related resourcing. This choice is defensible as a transparency–comparability trade-off, but it is not conceptually neutral. SG&A is a bundle that may capture selling and customer-facing effort, service and support capacity, administrative overhead and, particularly in compliance-intensive businesses, process, assurance and governance costs. Accordingly, treating SG&A expenses as equivalent to marketing investment risks construct contamination and biased interpretation [11].

A second, underappreciated issue is that different SG&A-derived measures operationalize different constructs and should not be treated as interchangeable. This study separates three indicators: commercialization efficiency (revenue/SG&A), commercialization intensity (SG&A/revenue) and SG&A growth (year-over-year change). Efficiency is a productivity-type indicator that approximates output per unit of a broader commercialization-related operating bundle. Intensity is a burden-type indicator that approximates the cost share required to sustain a given revenue level. Although mechanically related, these ratios are conceptually distinct: they correspond to different managerial questions (productivity versus cost burden) and may move in opposite directions across periods or scaling regimes.

Beyond construct validity, SG&A-based proxies are exposed to measurement error via managerial reporting choices and cross-firm classification heterogeneity. Evidence indicates that the apparent properties of SG&A can be distorted by classification shifting, with SG&A reclassified into other expense categories in ways that change how analysts and researchers perceive operating cost behavior [24]. For commercialization metrics, this implies that SG&A movements can reflect reporting incentives or accounting reclassification in addition to underlying shifts in customer-facing effort or operational capacity. The implication is not that SG&A expenses should be rejected, but that SG&A-derived measures should be interpreted as indicators of a mixed operating bundle rather than as precise marketing spend measures.

A further interpretability constraint concerns cost adjustment frictions. SG&A often behaves asymmetrically with respect to activity changes: it may not contract proportionately when activity declines and may adjust with delay when conditions change. A systematic review of cost stickiness research emphasizes both the prevalence of asymmetric cost behavior and the diversity of alignments across contexts, highlighting why SG&A dynamics cannot be assumed to be contemporaneously and linearly responsive to revenue movements [14]. This supports a disciplined reading of SG&A growth: increases may reflect capacity commitments, capability-building investments or delayed adjustment rather than immediate incremental commercialization activity [12].

Beyond SG&A constructs, revenue scale in CeFi is plausibly tied to the scale of operating capacity. Unlike purely digital products with near-zero marginal costs, CeFi intermediation requires ongoing investments in compliance operations, risk management, custody and security controls, customer support and transaction processing resilience. As activity expands, these functions often scale with volume and regulatory load, creating a co-scaling pattern between operating expenses and revenues. Digital transformation evidence in financial services highlights that performance gains are contingent on the cost base required to build and operate digital capabilities. As a result, growth may be accompanied by material operating expense expansion rather than costless scaling [8,13]. This supports examining operating expenses as a sector-level capacity indicator.

Taken together, these studies motivate triangulation rather than reliance on a single proxy. Using efficiency (revenue/SG&A), intensity (SG&A/revenue) and SG&A growth allows the analysis to distinguish productivity-type scaling patterns from cost-share pressure and potential adjustment dynamics while maintaining interpretive discipline. Accordingly, SG&A operationalization is used as an auditable, comparable signal of commercialization-related operating posture, useful for benchmarking and scenario reasoning, while remaining explicit about what it cannot identify [10].

Importantly, commercialization efficiency (revenue/SG&A) and commercialization intensity (SG&A/revenue) are mathematically related but economically distinct constructs. Efficiency captures output conversion per unit of the SG&A-based operating bundle, whereas intensity captures the relative cost share required to sustain revenue. Empirical marketing research increasingly emphasizes that productivity-oriented metrics may yield more decision-relevant insights than expenditure levels alone, particularly when input measures are noisy or aggregated [9,10]. Distinguishing these constructs is therefore not merely algebraic but central to interpreting whether commercialization contributes to growth through improved conversion or constrains profitability through cost pressure.

2.4. Marketing Efficiency, Growth–Margin Trade-Offs and Intertemporal Effects

Building on the SG&A-based constructs discussed above, commercialization productivity (revenue/SG&A) and commercialization intensity (SG&A/revenue) may exhibit distinct relationships with growth and profitability. Commercialization productivity (revenue/SG&A) reflects the conversion of bundled commercialization-related resources into output, whereas commercialization density or intensity (SG&A/revenue) reflects the structural weight of such expenditures relative to realized revenue. Marketing performance assessment research differentiates between input–output conversion efficiency and cost structure burden when evaluating accountability and financial impact [9,10]. These dimensions should not be conflated, as productivity shifts and cost-density shifts may exhibit distinct relationships with growth and profitability.

Marketing efficiency is typically framed as the ability to generate revenue outcomes relative to marketing-related resource commitments, with emphasis on decision usefulness rather than metric abundance [10]. In digital contexts, however, efficiency is not purely contemporaneous: marketing-related investments can create intangible assets (e.g., brand capital, customer relationships, trust and familiarity) whose payoffs accrue with delay, while accounting recognition largely expenses these investments, mechanically depressing period margins even when longer-run performance improves [25]. This time profile is central for interpreting a growth–profitability tension: commercialization can support revenue expansion while compressing near-term margins because costs are recognized immediately whereas benefits may accumulate gradually.

Observed returns are also sensitive to measurement and attribution adequacy. In digital advertising markets, attribution limitations and platform constraints can bias inferences about effectiveness and ROI, encouraging misallocation and noisy performance evaluation even when data appear abundant [26]. Accordingly, empirical associations involving efficiency constructs should be treated as performance-relevant signals rather than definitive causal estimates, especially when inputs are proxied using accounting aggregates instead of campaign-level spending data [10,26].

Intertemporal response dynamics further complicate efficiency assessment. Advertising response evidence shows that effects can reverse over time, implying that contemporaneous performance metrics may understate deferred benefits or obscure lagged channels [27]. While the empirical context differs, the underlying point generalizes that marketing-related spending can have immediate cost consequences and delayed revenue consequences, so observed efficiency–performance relationships depend on the time window and whether outcomes are defined as growth, profitability or scale.

In addition, marketing expenditures and sales can be jointly determined. Econometric work in B2B settings documents reciprocal relationships between sales and marketing expenses, consistent with the reality that budgets respond to realized sales performance while also shaping subsequent trajectories [28]. For archival designs, this bidirectionality implies that observed links between commercialization-related expenses and performance may reflect feedback and reactive budgeting as well as proactive investment effects.

Recent evidence also cautions that naive contemporaneous analyses can make marketing appear unprofitable, whereas incorporating lag structures and nonlinearities can change inferred profitability relationships and reveal over- or under-spending regimes [29]. For the present study, these insights motivate separating growth-oriented from profitability-oriented outcomes and allowing for intertemporal patterns when examining SG&A dynamics rather than assuming immediate and uniform payoffs [25,27,29].

2.5. From Empirical Associations to Decision Support: FCM-Based Scenario Modeling

Regression analysis can establish whether SG&A-proxied commercialization indicators are systematically associated with revenue growth and profitability, but it is less suited to managerial contexts where multiple levers shift simultaneously. Because CeFi operating structures and commercialization posture often co-move, decision-makers require structured “what-if” reasoning about joint shocks and feedback-consistent responses rather than single-variable marginal interpretation. Fuzzy cognitive maps (FCMs) support scenario exploration in systems characterized by interdependence and feedback, enabling semi-quantitative propagation of changes across linked constructs [15,17,30].

The FCM layer is positioned as a complement to the econometric analysis, not as a replacement or a causal identification device. Its value lies in making the assumed linkage structure explicit and inspectable and in enabling sensitivity analysis under alternative perturbations, an advantage emphasized in recent syntheses of FCM use in complex decision settings [16]. At the same time, methodological reviews emphasize that outputs are contingent on the specified structure and weights, so scenario results should be interpreted as directional insights rather than forecasts [31]. Accordingly, Section 4.4 documents the FCM specification, weight construction, transfer function choice and scenario design.

For managers in CeFi institutions, commercialization decisions rarely occur in isolation; changes in SG&A allocation often coincide with adjustments in operating capacity, compliance infrastructure and digital service investment. The FCM layer therefore operationalizes the regression sign structure into a bounded dynamic system that allows exploration of joint shocks rather than single-coefficient interpretation. The objective is not forecasting precision but structured sensitivity analysis consistent with decision-support system research in digital environments [15,32].

Taken together, the reviewed literature clarifies mechanisms linking digital capabilities, marketing accountability, cost behavior and performance outcomes. However, it does not resolve whether SG&A-based commercialization indicators retain explanatory power in centralized financial intermediation, whether productivity and intensity dimensions exert differentiated effects under aggregation constraints or how such empirical associations can be translated into structured scenario reasoning for managerial use. These unresolved issues define the analytical focus of the present study.

2.6. Conceptual Framework

Figure 1 summarizes the study’s conceptual framework and links the theorized information-enabled commercialization posture to the accounting-based constructs used in the empirical analysis. Because direct, comparable measures of IS capability maturity and marketing system sophistication are not consistently observable across CeFi firms, the framework treats financial statement aggregates as auditable signals of a bundled commercialization–operations posture in which market-facing effort, service delivery and governance execution are intertwined. Within this bundled posture, the framework distinguishes four analytically separable dimensions: (i) commercialization efficiency (revenue/SG&A), capturing the conversion/productivity of the bundle, (ii) commercialization intensity (SG&A/revenue), capturing the structural cost burden, (iii) SG&A growth, capturing expansion pace and adjustment dynamics and (iv) operating expense scale, capturing the extent of operating capacity required to support activity.

The hypotheses map these dimensions to distinct performance outcomes. Efficiency is expected to align positively with revenue growth (H1), whereas intensity is expected to align negatively with net margin (H2). Operating expense scale is expected to co-move positively with revenue scale (H3), reflecting the capacity requirements of compliance-, risk- and service-facing functions in CeFi. Finally, SG&A growth is expected to exhibit intertemporal patterns relative to subsequent growth (H4), consistent with adjustment frictions and delayed payoffs. The fuzzy cognitive map is positioned as a complementary decision-support representation that makes assumed linkages explicit and supports scenario reasoning. In this study, information-enabled refers to commercialization executed through auditable information and governed processes, as reflected in accounting aggregates and scenario logic, not direct observation of IS capability maturity.

3. Materials and Methods

3.1. Methodology

•

Stage 1: Data selection and gathering

This study uses a reproducible, aggregated dataset that captures the financial and commercialization trajectory of the CeFi sector over 2020–2025. The unit of analysis is the annual aggregated sector level, constructed from a predefined list of CeFi firms (reported earlier in the study) and harmonized into a single time series. For each year, core financial statement variables were gathered and aggregated, including revenue, net income, total assets, total liabilities, total equity, SG&A, and operating expenses (all in USD millions). These inputs were then used to compute standardized ratio measures (e.g., SG&A intensity, operating expense intensity, and revenue-per-cost efficiency ratios) and growth rates (YoY changes), ensuring comparability across years. The Supplementary Material contains the dataset used in the analysis, including annual financial variables (revenue, net income, assets, liabilities, equity), SG&A and operating expenses, and derived financial ratios and growth metrics (accessed on 18 January 2026).

SG&A is an inherently bundled accounting category incorporating heterogeneous components such as marketing, administrative, support, and compliance-related expenditures. Consequently, SG&A-based measures do not isolate pure marketing investment but reflect a broader commercialization–operation cost structure. This aggregation may introduce construct contamination, potentially attenuating or distorting estimated associations depending on cost composition, reporting practices, and adjustment dynamics.

To improve transparency and replicability, variable construction follows deterministic formulas (e.g., sga_pct_revenue = sga_musd/revenue_musd) and all transformations are documented. Growth metrics are computed as year-over-year percentage changes, which reduces sensitivity to scale and supports longitudinal interpretation. The final dataset consists of six annual observations (2020–2025), with growth variables available for five observations due to lag requirements.

•

Stage 2: Descriptive statistics

The second stage characterizes the dataset using descriptive statistics to establish the distributional properties of each variable and provide context for subsequent modeling. For each measure, the study reports N, mean, standard deviation, minimum, and maximum. This stage serves two purposes. First, it documents the magnitude and dispersion of CeFi financial outcomes and cost structures across the 2020–2025 period. Second, it helps identify potentially influential observations and the extent of variability in commercialization intensity and efficiency proxies, which is essential for interpreting correlations and regression coefficients in a short time series.

•

Stage 3: Statistical analysis and regression models

The third stage assesses linear associations and tests using IBM SPSS Statistics (31.0.1.0) the proposed mechanisms using correlation analysis and parsimonious OLS regression models. The dataset consists of a short aggregated annual time series (2020–2025, N = 6). Accordingly, the OLS estimations are treated as exploratory association models designed to detect directional patterns rather than to support population-level inference or predictive generalization. Pearson correlations are computed for the complete set of variables to establish baseline relationships between scale measures (e.g., revenue and assets), cost structure (SG&A and opex), intensity ratios (SG&A/revenue and opex/revenue), efficiency ratios (revenue/SG&A and revenue/opex), profitability (net margin), and growth dynamics (YoY measures). Given the strong co-movement among balance-sheet totals (e.g., assets and liabilities) and among ratio constructs (e.g., opex intensity vs. revenue per opex), the analysis emphasizes model parsimony to mitigate multicollinearity.

Four OLS models are estimated to align directly with the hypotheses: (M1) revenue growth explained by SG&A efficiency and SG&A growth; (M2) profitability explained by SG&A intensity and leverage; (M3) revenue scale explained by operating expense scale in log–log form; and (M4) revenue growth explained by lagged SG&A growth. Coefficients are reported with standard errors and conventional significance levels, and model fit statistics (R², adjusted R², F-test, and p-values) are presented for transparency. The empirical design does not permit causal inference. The OLS estimations identify conditional associations within the observed aggregated time series and do not establish causal direction, as reciprocal relationships and omitted variable bias remain plausible.

•

Stage 4: Fuzzy cognitive map modeling

The fourth stage extends the regression-based findings into a systems perspective using fuzzy cognitive map (FCM) using Mental Modeler (version 14.0). FCM is employed to represent CeFi performance as a causal network of interacting constructs—commercial investment (SG&A), operational capacity (opex), commercialization efficiency (revenue-per-SG&A), cost intensity (SG&A/revenue and opex/revenue), financial structure (assets, liabilities, equity), profitability (net margin), and growth (revenue YoY). Nodes correspond to measured variables (or constructs derived from them), while directed edges represent causal influences informed by the empirical signs and magnitudes observed in the correlation and regression stages.

FCM development proceeds by (i) defining the concept set and map variables to concepts, (ii) assigning initial edge weights based on standardized regression coefficients and/or strong correlation evidence (with explicit rules for translating statistical effects into fuzzy weights), and (iii) running scenario simulations to evaluate how changes in key controllable nodes (e.g., improving SG&A efficiency or reducing cost intensity) propagate through the system to affect growth and profitability. This stage enables “what-if” analysis consistent with the journal’s information systems scope, translating statistical associations into an interpretable decision model that supports reproducible experimentation and sensitivity analysis.

3.2. Sample Retrieval

The sample was retrieved through a transparent, reproducible selection process targeting CeFi firms with consistent, publicly available financial disclosures over the study window (2020–2025). First, an initial universe of large CeFi organizations was defined to represent centralized financial intermediation and digital financial services, including major banking institutions and publicly listed FinTech/CeFi platforms (JPMorgan Chase & Co., Bank of America, Wells Fargo, Goldman Sachs, Morgan Stanley, PayPal, Block, Inc., Coinbase, Robinhood Markets, and SoFi Technologies). Firms were retained if they provided continuous annual reporting for the required income statement and balance-sheet variables (revenue, net income, total assets, total liabilities, total equity, SG&A, and operating expenses) across the observation period (

Table 1. Variable definitions and measurements.

Variable	Construct	Definition (Plain)	Formula	Unit/Scale	Expected Role in Your Study
revenue_musd	Financial performance (scale)	Total revenue aggregated across CeFi firms	Reported value	USD millions	Scale indicator; also a denominator for intensity ratios; used for growth
net_income_ musd	Financial performance (profit)	Net income aggregated across CeFi firms	Reported value	USD millions	Profit level; numerator for margin
total_assets_ musd	Financial structure (size/capacity)	Total assets aggregated across CeFi firms	Reported value	USD millions	Size/control variable; capacity/resilience proxy
total_ liabilities_musd	Financial structure (leverage)	Total liabilities aggregated across CeFi firms	Reported value	USD millions	Balance-sheet risk/funding proxy; used to derive leverage measures
total_equity_musd	Financial structure (capital buffer)	Total equity aggregated across CeFi firms	Reported value	USD millions	Capital strength proxy; used to derive equity ratio
sga_musd	Marketing/ commercial investment (proxy)	SG&A expense (includes sales, marketing, admin), used as a proxy for commercialization effort	Reported value	USD millions	Marketing/commercial input; basis for intensity/efficiency metrics
operating_ expenses_musd	Operating efficiency (scale)	Total operating expenses aggregated across CeFi firms	Reported value	USD millions	Operations/platform cost base; basis for opex intensity/efficiency
sga_pct_revenue	Marketing intensity	SG&A as a share of revenue (how “heavy” commercial/admin costs are)	sga_musd/revenue_musd	Ratio (0–1)	Marketing/commercial cost intensity predictor (often expected ↓lower margin)
opex_pct_ revenue	Operating cost intensity	Operating expenses as a share of revenue	operating_expenses_musd/revenue_musd	Ratio (0–1)	Operational efficiency predictor (higher intensity often ↓lower margin)
rev_per_sga	Marketing efficiency	Revenue generated per unit of SG&A	revenue_musd/sga_musd	Ratio	Commercial efficiency predictor (often expected ↑ higher growth, ↑higher margin)
rev_per_opex	Operational efficiency	Revenue generated per unit of operating expenses	revenue_musd/operating_expenses_musd	Ratio	Operational efficiency predictor (often expected ↑higher margin)
net_margin	Profitability	Net income as a share of revenue	net_income_musd/revenue_musd	Ratio	Key dependent variable for profitability-focused hypotheses
sga_share_opex	Cost structure allocation	Portion of operating expenses represented by SG&A	sga_musd/operating_expenses_musd	Ratio (0–1)	Structure/strategy proxy: how much opex is commercial/admin vs. other ops
revenue_yoy_ growth	Growth	Year-over-year revenue growth	(revenue_t − revenue_(t − 1))/revenue_(t − 1)	Rate	Key dependent variable for growth-focused hypotheses
sga_yoy_growth	Marketing investment dynamics	Year-over-year SG&A growth	(sga_t − sga_(t − 1))/sga_(t − 1)	Rate	Key independent variable (investment expansion), also used with lags

). For each year, firm-level values were extracted and harmonized to a common unit (USD millions), then aggregated to produce the annual sector-level series used in the analysis. Derived measures, cost intensities, efficiency ratios, and year-over-year growth rates were computed deterministically from these aggregated totals using documented formulas, ensuring that the final dataset can be reconstructed from the same sources under identical inclusion criteria.

3.3. Research Hypotheses

Building on the literature linking information-enabled commercialization capabilities to firm performance, this study formulates four hypotheses that connect SG&A-based marketing proxies and operating cost structure to CeFi financial outcomes. Specifically, we test whether commercialization efficiency and investment dynamics explain revenue growth, whether commercialization intensity constrains profitability when controlling for leverage, whether operating expense scale tracks revenue scale, and whether rapid SG&A expansion exhibits lagged “payback/overshoot” effects in subsequent growth.

H1: 

Higher SG&A efficiency and higher SG&A growth are associated with higher revenue growth.

This hypothesis draws on the marketing productivity and accountability literature, which conceptualizes marketing performance as the conversion of inputs (commercial spending, customer communication, and selling efforts) into outcomes (sales, growth, and cash flows). Marketing efficiency—captured here as revenue generated per unit of SG&A—fits directly within established frameworks that distinguish marketing efficiency from effectiveness and emphasize linking marketing inputs to financial outcomes [33,34,35].

Empirically, marketing capabilities and commercialization-related investments have been shown to explain variance in revenue growth and profit growth components, supporting the view that better “commercial execution” (often enabled by CRM, analytics, and brand management) is associated with superior growth performance. In our setting, SG&A is used as a pragmatic proxy for commercialization investment when direct marketing expenditure or digital marketing platform metrics are not uniformly available. The literature also cautions that SG&A is an imperfect proxy for marketing spend and can lead to biased inferences, which strengthens the motivation to focus on efficiency (output per SG&A) rather than SG&A levels alone [11,36].

H2: 

Higher SG&A intensity (SG&A as a share of revenue) is associated with lower profitability, controlling for leverage.

SG&A intensity captures the “cost burden” of commercialization and administrative infrastructure relative to revenue. Cost-to-sales ratios are widely used in performance diagnostics because they reflect whether the organization’s selling/administrative structure is scaled efficiently relative to realized demand. Consistent with this logic, recent empirical evidence shows that higher SG&A expense ratios can be associated with lower operating profit margins, emphasizing cost discipline and resource allocation efficiency [37].

Leverage control is theoretically important because capital structure can shape managerial constraints and spending flexibility, potentially influencing discretionary cost lines such as SG&A expenses. In finance/accounting research, leverage is often linked to incentives or constraints that can affect real decisions on discretionary expenditures (including SG&A adjustments), which is why controlling for leverage helps isolate the profitability relationship tied to SG&A intensity rather than balance-sheet pressure [12,38].

H3: 

Higher operating expense scale is positively associated with revenue scale.

This hypothesis follows from scaling and economies-of-scale logic: as service platforms expand output, operating expenses (staffing, technology operations, processing, customer support, and compliance infrastructure) tend to scale with revenue generation capacity. Classic evidence in financial institutions documents operating cost patterns and economies of scale in different expense categories, motivating the expectation that larger operational footprints correspond to larger revenue bases [39,40].

From an information systems lens, operating expense scale can also be interpreted as a proxy for information processing capacity (digitalization, automation, and service infrastructure). The financial sector has experienced extensive digitalization and structural changes in operating cost composition, reinforcing the relevance of examining how operating cost bases track sector revenue capacity over time [41,42].

H4: 

Higher prior-year SG&A growth is associated with lower subsequent revenue growth (consistent with overspending/payback dynamics).

Marketing and commercialization investments frequently exhibit dynamic effects: outcomes may appear with delay (carryover) and may also show diminishing returns at higher spending levels. Advertising response and long-term marketing effectiveness research documents carryover and persistence mechanisms, implying that spending changes can affect future outcomes, not only contemporaneous ones [43,44].

At the same time, the literature highlights diminishing returns and the risk of inefficiency or overspending—especially when managers rely on noisy performance metrics—creating a plausible “overshoot/payback” pattern where unusually high spending growth is followed by weaker subsequent growth. This supports lagged negative association as an efficiency-correction mechanism: aggressive SG&A expansion may reflect front-loaded acquisition/expansion costs, misallocation, or saturation effects that depress the next period’s growth rate [45,46,47].

4. Results

4.1. Descriptive Statistics

Table 2. Descriptive statistics.

Variable	N	Mean	Std. Dev.	Min	Max
revenue_musd	6	47,011.0772	6594.8235	36,005.9530	54,531.6000
net_income_musd	6	12,230.0995	3101.1750	7537.7380	15,568.3240
total_assets_musd	6	1,189,743.1000	96,117.1857	1,054,184.0000	1,343,671.5000
total_liabilities_musd	6	1,083,695.4333	89,301.3829	957,364.1000	1,227,749.6000
total_equity_musd	6	106,048.3833	7115.0909	96,820.0000	115,921.9000
sga_musd	6	11,556.7100	1202.2919	9991.4267	13,342.1667
operating_expenses_musd	6	12,819.4518	1800.8945	9680.7100	14,895.2857
sga_pct_revenue	6	0.2132	0.0321	0.1758	0.2698
opex_pct_revenue	6	0.7075	0.1401	0.5411	0.9423
rev_per_sga	6	5.2867	0.3964	4.8002	5.8481
rev_per_opex	6	1.9567	0.2457	1.6011	2.1796
net_margin	6	0.1144	0.1443	−0.0492	0.2866
sga_share_opex	6	0.5459	0.0205	0.5115	0.5654
revenue_yoy_growth	5	0.2745	0.3888	−0.0773	0.9404
sga_yoy_growth	5	0.2067	0.3093	−0.0381	0.7325

summarizes the aggregated CeFi trajectory for 2020–2025. Average revenue is 47,011.08 million USD (SD = 6594.82), while net income averages 12,230.10 million USD (SD = 3101.18), indicating substantial year-to-year variability in both scale and profitability. Balance-sheet size is large and comparatively stable: total assets average 1,189,743.10 million USD and total equity averages 106,048.38 million USD, suggesting that structural capitalization changes more gradually than income statement performance.

Commercial and operating costs show meaningful dispersion. SG&A averages 11,556.71 million USD and operating expenses 12,819.45 million USD, with operating expenses exhibiting higher variability (SD = 1800.89). SG&A intensity (sga_pct_revenue) averages 0.2132 (≈21% of revenue), while operating expense intensity (opex_pct_revenue) averages 0.7075, reflecting a heavy cost base relative to revenue that varies across the period.

Beyond documenting dispersion, Table 2 directly reflects the measurement and construct challenges outlined in Section 2. The coexistence of substantial variability in SG&A intensity (mean 0.2132; range 0.1758–0.2698) and efficiency (rev_per_sga mean 5.2867) indicates that commercialization burden and commercialization productivity are not mechanically equivalent, despite being algebraically related. This supports the paper’s central differentiation between cost density and output conversion.

Furthermore, the heavy operating expense intensity (mean 0.7075) suggests that CeFi activity is embedded within a substantial structural cost base, consistent with the argument that centralized financial intermediation requires compliance, service, and infrastructure capacity beyond pure marketing effort. The volatility in revenue growth (range −0.0773 to 0.9404) combined with variability in SG&A growth reinforces the need to examine contemporaneous and lagged relationships rather than assuming linear proportionality between spending and performance. Thus, the descriptive statistics already illustrate the measurement gap motivating this study: SG&A is neither a clean marketing input nor a homogeneous cost category, but a bundled commercialization–operations signal requiring analytical decomposition.

4.2. Correlation Analysis

Table 3. Correlation matrix.

	Revenue_Musd	Net_Income_Musd	Total_Assets_Musd	Total_Liabilities_Musd	Total_Equity_Musd	Sga_Musd	Operating_Expenses_Musd	Sga_Pct_Revenue	Opex_Pct_Revenue	Rev_Per_Sga	Rev_Per_Opex	Net_Margin	Sga_Share_Opex	Revenue_Yoy_Growth	Sga_Yoy_Growth
revenue_musd	1.0000	0.9219	0.8875	0.8866	0.8782	0.9762	0.9844	−0.7183	−0.9628	0.7566	0.9595	0.8095	0.2748	0.3193	−0.2746
net_income_musd	0.9219	1.0000	0.9621	0.9627	0.9403	0.9638	0.9495	−0.7559	−0.9075	0.7940	0.9021	0.8148	0.1254	0.6202	0.0325
total_assets_musd	0.8875	0.9621	1.0000	0.9999	0.9873	0.9394	0.9134	−0.7811	−0.8538	0.7887	0.8491	0.7417	−0.0323	0.7347	0.2573
total_liabilities_musd	0.8866	0.9627	0.9999	1.0000	0.9861	0.9389	0.9127	−0.7815	−0.8530	0.7890	0.8484	0.7415	−0.0334	0.7350	0.2577
total_equity_musd	0.8782	0.9403	0.9873	0.9861	1.0000	0.9301	0.9025	−0.7961	−0.8362	0.8030	0.8288	0.7415	−0.0204	0.7414	0.2510
sga_musd	0.9762	0.9638	0.9394	0.9389	0.9301	1.0000	0.9900	−0.6799	−0.9362	0.7282	0.9352	0.7590	0.1451	0.4896	−0.1166
operating_expenses_musd	0.9844	0.9495	0.9134	0.9127	0.9025	0.9900	1.0000	−0.6559	−0.9587	0.7067	0.9629	0.7774	0.2570	0.3787	−0.2238
sga_pct_revenue	−0.7183	−0.7559	−0.7811	−0.7815	−0.7961	−0.6799	−0.6559	1.0000	0.5570	−0.9674	−0.5416	−0.8638	−0.3620	−0.7052	−0.4087
opex_pct_revenue	−0.9628	−0.9075	−0.8538	−0.8530	−0.8362	−0.9362	−0.9587	0.5570	1.0000	−0.6102	−0.9973	−0.7285	−0.3180	−0.1944	0.3795
rev_per_sga	0.7566	0.7940	0.7887	0.7890	0.8030	0.7282	0.7067	−0.9674	−0.6102	1.0000	0.5944	0.8848	0.2774	0.7764	0.4489
rev_per_opex	0.9595	0.9021	0.8491	0.8484	0.8288	0.9352	0.9629	−0.5416	−0.9973	0.5944	1.0000	0.7192	0.3260	0.1914	−0.3792
net_margin	0.8095	0.8148	0.7417	0.7415	0.7415	0.7590	0.7774	−0.8638	−0.7285	0.8848	0.7192	1.0000	0.4132	0.7517	0.2164
sga_share_opex	0.2748	0.1254	−0.0323	−0.0334	−0.0204	0.1451	0.2570	−0.3620	−0.3180	0.2774	0.3260	0.4132	1.0000	0.2380	0.5703
revenue_yoy_growth	0.3193	0.6202	0.7347	0.7350	0.7414	0.4896	0.3787	−0.7052	−0.1944	0.7764	0.1914	0.7517	0.2380	1.0000	0.9124
sga_yoy_growth	−0.2746	0.0325	0.2573	0.2577	0.2510	−0.1166	−0.2238	−0.4087	0.3795	0.4489	−0.3792	0.2164	0.5703	0.9124	1.0000

indicates strong co-movement between financial scale and cost structure. Revenue is highly correlated with SG&A (r = 0.9762) and operating expenses (r = 0.9844), and it also correlates strongly with net income (r = 0.9219). This suggests that expansion in the aggregated CeFi sector is accompanied by proportional increases in commercialization and operational spending.

Intensity measures behave as cost burden indicators. SG&A intensity (sga_pct_revenue) is strongly negatively correlated with net margin (r = −0.8638) and revenue (r = −0.7183), while operating expense intensity (opex_pct_revenue) shows similarly strong negative associations with scale and efficiency (e.g., with rev_per_opex, r = −0.9973). Conversely, efficiency measures correlate positively with performance: rev_per_sga correlates with net margin (r = 0.8848) and revenue growth (r = 0.7764).

Several near-perfect correlations (e.g., assets with liabilities, r = 0.9999) imply multicollinearity risk, supporting a parsimonious regression strategy rather than including many predictors simultaneously. Growth measures also move together: revenue YoY growth correlates strongly with SG&A YoY growth (r = 0.9124), motivating both contemporaneous and lagged spending tests in the regressions.

Importantly, the correlation structure illustrates why construct differentiation is necessary. The strong positive association between revenue and SG&A (r = 0.9762) could superficially suggest that commercialization spending “drives” growth; however, the simultaneous strong negative correlation between SG&A intensity and net margin (r = −0.8638) indicates that cost burden and revenue scale encode different financial implications. This directly addresses the measurement gap identified earlier: treating SG&A as a single marketing proxy obscures whether observed associations reflect productivity, structural cost burden, or scaling effects.

The near-perfect correlations among balance-sheet aggregates further justify the parsimonious regression strategy and highlight that sector-level CeFi expansion is structurally coupled rather than composed of independent financial movements. This structural coupling motivates the later FCM representation, where interdependencies are modeled explicitly rather than interpreted as isolated pairwise relationships.

4.3. Regression Analysis

Table 4. OLS coefficients.

Term	M1 RevGrowth	M2 NetMargin	M3 Log(Rev)	M4 RevGrowth Lag
Intercept	−4.529 ** (0.647)	0.249 (0.399)	1.562 (1.348)	0.201 ** (0.030)
SGA efficiency (rev_per_sga_from_totals)	1.131 ** (0.160)
SGA YoY growth (sga_yoy_growth)	0.689 ** (0.102)
SGA intensity (sga_pct_revenue_from_totals)		−1.501 * (0.489)
Leverage (liab/equity)		0.037 (0.031)
log(Opex)			0.972 *** (0.143)
Lagged SGA YoY growth (sga_yoy_lag1)				−0.397 ** (0.079)

Significance: *** p < 0.01, ** p < 0.05, * p < 0.10.

provides OLS results consistent with the study’s marketing efficiency mechanism. High R² values are expected in short time series regressions with strongly trending variables and should not be interpreted as evidence of predictive robustness or absence of overfitting. Moreover, variations in control variable coefficients across models are expected, as each specification conditions on different predictor sets and targets distinct dependent variables. Given the short aggregated time series, such coefficient sensitivity should not be interpreted as structural parameter shifts but as model-specific adjustments.

In M1, revenue growth is positively associated with SG&A efficiency (rev_per_sga_from_totals, β = 1.131, p < 0.05) and SG&A growth (sga_yoy_growth, β = 0.689, p < 0.05), with very high model fit (R² = 0.994, N = 5). This indicates that commercialization efficiency and spending expansion are closely aligned with aggregated revenue growth over the period.

In M2, profitability is negatively associated with SG&A intensity (sga_pct_revenue_from_totals, β = −1.501, p < 0.10), while leverage is not statistically significant. This supports the interpretation that when SG&A consumes a larger share of revenue, margins weaken, even after accounting for capital structure at the aggregate level (R² = 0.885, N = 6). In M3, operating expense scale predicts revenue scale strongly (log–log elasticity β = 0.972, p < 0.01), consistent with proportional scaling between operating capacity and revenue generation (R² = 0.921, N = 6).

Finally, M4 shows a negative lag relationship: prior SG&A growth predicts lower subsequent revenue growth (sga_yoy_lag1, β = −0.397, p < 0.05; R² = 0.926, N = 4). This is consistent with a potential overshoot or delayed-payback pattern in commercialization spending. Given the small annual sample and reduced N in the lag model, this evidence is best interpreted as suggestive and suitable for validation using firm-level panels in future work. Below, the four models (M1–M4) of the developed regressions are presented:

•

M1: revenue_yoy_growth_t=α+β₁rev_per_sga_from_totals_t+β₂sga_yoy_growth_t+ε_t.

•

M2: net_margin_from_totals_t=α+β₁sga_pct_revenue_from_totals_t+β₂leverage_t+ε_t.

•

M3: ln(revenue_musd_t)=α+β₁ln(operating_expenses_musd_t)+ε_t.

•

M4: revenue_yoy_growth_t=α+β₁sga_yoy_lag1_t+ε_t.

4.4. FCM Simulation

The FCM layer translates the sign structure documented in Table 3, Table 4 and

Table 5. Model fit statistics.

Model	N	R2	Adj. R2	F	Prob(F)
M1 RevGrowth	5	0.994	0.987	153.298	0.006
M2 NetMargin	6	0.885	0.808	11.513	0.039
M3 log(Rev)	6	0.921	0.901	46.467	0.002
M4 RevGrowth lag	4	0.926	0.889	25.058	0.038

into a system-level representation that makes interdependencies explicit. While the regression models identify conditional associations within the short aggregated time series, the FCM operationalizes the research gap concerning decision-support translation. By embedding efficiency, intensity, scale, and growth nodes within a bounded dynamic network, the model demonstrates how the differentiated commercialization constructs identified in the statistical tables propagate jointly under coordinated shocks. The simulation results therefore extend, rather than replicate, the regression findings by illustrating systemic sensitivity consistent with the measurement and scaling patterns observed empirically.

Section 4.4 extends the statistical findings with a fuzzy cognitive map (FCM) simulation to evaluate how CeFi financial and commercialization variables interact as a coupled system under alternative conditions. While correlation and OLS models quantify pairwise and conditional relationships, FCM modeling enables dynamic “what-if” analysis by propagating shocks across interconnected constructs (e.g., SG&A dynamics, operating cost structure, revenue scale, and profitability) until the system reaches a stable state. In this way, the simulation provides an interpretable, reproducible mechanism for assessing how simultaneous changes in key drivers may translate into changes in revenue growth, Net Margin from Totals, and Total Revenues.

Figure 2 illustrates the developed fuzzy cognitive map (FCM) that operationalizes the CeFi performance system as a network of interacting financial and marketing-related concepts. In line with the foundational definition of FCMs as fuzzy, directed causal graphs that support causal propagation and feedback, each node represents a measurable construct (e.g., revenue growth, Net Margin from Totals, Total Revenues, SG&A intensity/efficiency, operating cost structure, and balance-sheet totals), while each directed edge captures the sign of influence (positive/negative) between concepts. This representation is particularly suitable for domains where multiple mechanisms operate simultaneously and feedback is plausible—conditions that characterize financial performance systems in digital finance [48,49,50].

The model development procedure follows established FCM practice: (i) concept selection grounded in the study variables, (ii) structural linking based on the empirical association patterns (correlation and regression signs), and (iii) weight initialization through a normalization rule so that stronger empirical relationships correspond to higher absolute weights. This provides a defensible “data-informed” FCM rather than a purely expert-elicited map, while still preserving the FCM advantage of representing system-wide interdependence rather than isolated bivariate relationships. Methodologically, this aligns with the broader FCM literature emphasizing that maps can be built from mixed evidence sources (data + theory) and then used as computational decision models [49,51].

For simulation, the FCM uses an iterative state-update process with a bounded activation/transfer function. The selection of the hyperbolic tangent is consistent with common FCM implementations because it constrains concept activation within a stable interval (typically −1 to +1), reducing the risk of explosive cycles when feedback loops exist and enabling convergence to interpretable steady states. Recent methodological work explicitly notes that sigmoid and hyperbolic tangent transfer functions are standard choices to stabilize the iterative procedure and prevent unrealistic extreme states [32,52,53].

Scenarios 1–3 (positive shocks) show a coherent expansion pathway. In the first three experiments, the intervention variables—Operating Expenses Total and SG&A growth—are increased in equal steps (+0.25, +0.50, +0.75). The predicted outcomes rise monotonically across all three target nodes: revenue growth increases from ~0.14 to ~0.18 and ~0.21, Net Margin from Totals increases from ~0.17 to ~0.20 and ~0.21, and Total Revenues increase from ~0.22 to ~0.28 and ~0.30. This pattern is consistent with the empirical results that (a) commercialization expansion and SG&A dynamics are positively related to growth (Model M1), and (b) operating cost scale tracks revenue scale (Model M3). In FCM terms, the network contains enough positive pathways from capacity/investment nodes to performance nodes that the system converges to higher equilibrium states when “inputs” rise.

The results also imply that “scale + commercialization momentum” is reinforced through interconnected mechanisms. The joint increase in Operating Expenses Total and SG&A growth activates multiple mediated links, e.g., operating scale → revenue scale, commercialization momentum → growth, and margin improvement through efficiency-related nodes—so the total effect observed in Figure 3 is the net outcome of several reinforcing paths rather than a single direct edge. This is one of the core advantages of FCM simulation versus static regression: it can illustrate how a policy or investment move propagates through a causal network with simultaneous interactions and feedback.

Scenarios 4–6 (negative shocks) produce systematic contraction effects. When the same two intervention variables are reduced (−0.25, −0.50, and −0.75), all target outcomes move in the opposite direction: revenue growth, Net Margin, and Total Revenues shift to negative state values and become more adverse as the shock magnitude increases (e.g., revenue growth to around −1.48 to −1.55; margin to around −1.74 to −1.78; revenues to around −1.61 to −1.68 in the displayed scale). This directional symmetry indicates that the modeled system is sensitive to reductions in operating capacity and commercialization momentum, a plausible property of digital finance systems where cutting operational capacity and commercial effort can rapidly affect acquisition, retention, and monetization. The strong propagation of negative shocks is also typical in feedback-rich FCMs, where a decline in one component can dampen others through multiple outgoing paths.

Finally, the pattern across scenarios is consistent with bounded nonlinearity and saturation. Because the simulation uses a hyperbolic tangent transfer function, increases in inputs tend to yield diminishing incremental gains as states approach the function’s bounds, while negative shocks can also compress states toward the lower bound. This makes the response curves in Figure 3 interpretable as nonlinear but stable, which is desirable for scenario work (it avoids unrealistic runaway growth/decline in iterative updates). Methodologically, this supports presenting the FCM scenarios as a robust “stress-test” layer that complements the regression evidence: regressions quantify linear associations, while the FCM demonstrates how combined changes in key drivers can shift the system-wide equilibrium under bounded dynamics [54,55]. The FCM simulations do not constitute independent empirical validation but rather a structural scenario representation based on the sign and directional patterns observed in the statistical analysis. Scenario outputs are therefore interpreted as illustrative and directional, not predictive.

5. Discussion

The regression results (Table 4) directly address the three research gaps articulated in the Introduction: the construct differentiation gap, the capacity scaling gap, and the decision-support gap. Rather than demonstrating a uniform “marketing spend effect,” Models M1–M4 reveal differentiated mechanisms embedded within SG&A-based proxies. Efficiency (revenue/SG&A) aligns positively with growth, intensity (SG&A/revenue) aligns negatively with margin, operating expense scale co-moves proportionally with revenue scale, and SG&A growth exhibits intertemporal adjustment dynamics. These findings confirm that SG&A-based commercialization indicators encode multiple analytically distinct relationships, validating the need to decompose rather than aggregate commercialization constructs under accounting constraints. H1 is supported by Model M1. Revenue growth is positively associated with SG&A-proxied commercialization efficiency (rev_per_sga_from_totals: β = 1.131, p < 0.05) and with contemporaneous SG&A growth (sga_yoy_growth: β = 0.689, p < 0.05). This result addresses the construct differentiation gap identified in the Introduction by showing that SG&A-based productivity measures capture commercialization conversion dynamics that are analytically distinct from expenditure levels alone. Substantively, this pattern is consistent with marketing accountability arguments that productivity-type metrics (output per unit of commercialization-related resource) can be more decision-useful than spend levels alone [10]. It also aligns with cautions in the SG&A-proxy literature: efficiency ratios may partially mitigate construct noise in SG&A aggregates because they foreground output conversion rather than treating SG&A levels as a clean “marketing input” [11].

In the CeFi context, this association can be interpreted through an IS-enabled trust-production mechanism. Commercialization efficiency does not reflect advertising conversion alone, but the joint effectiveness of onboarding systems, risk controls, transaction reliability and customer support orchestration delivered through centralized digital infrastructures. Because CeFi intermediaries retain custody and regulatory responsibility, user acquisition and retention are closely tied to perceived operational reliability and compliance credibility rather than communication exposure alone [1,18]. Higher revenue per SG&A may therefore signal not only effective customer acquisition but also scalable trust execution embedded in information systems and governance processes. This interpretation is consistent with IS capability research showing that performance gains often arise through process execution quality rather than through isolated front-end investments [6].

H2 is directionally supported by Model M2. Net margin is negatively associated with SG&A intensity (sga_pct_revenue_from_totals: β = −1.501, p < 0.10), while leverage is not statistically significant. The sign is consistent with interpreting intensity as a cost burden ratio: when SG&A consumes a larger share of revenue, margin pressure increases. This finding further supports the construct differentiation argument by demonstrating that SG&A-derived indicators capture different economic mechanisms: while efficiency relates to growth outcomes, intensity reflects structural cost burden affecting profitability. However, the weaker significance threshold and short series warrant restraint. Because SG&A reflects a mixed operating bundle that includes compliance and administrative components, intensity effects on profitability should be interpreted cautiously [11,12].

In regulated centralized finance, elevated SG&A intensity may reflect structural compliance and governance requirements rather than discretionary marketing overspending. Anti-money laundering procedures, KYC infrastructure, dispute resolution, audit processes and regulatory reporting systems are typically embedded within administrative expense lines. As revenue growth moderates, these partially fixed or adjustment-friction-sensitive costs may exert disproportionate pressure on margins. Cost stickiness research suggests that SG&A components often adjust asymmetrically to demand contractions, reinforcing margin compression during slower growth phases [12,14]. Accordingly, the negative intensity–margin relationship may capture regulatory and structural cost burdens intrinsic to centralized intermediation rather than inefficient commercialization alone.

H3 is supported by Model M3. The log–log elasticity of revenue with respect to operating expenses is positive and statistically significant (log(Opex): β = 0.972, p < 0.01). This indicates strong co-movement between the aggregated cost base and revenue scale over 2020–2025, consistent with a capacity co-scaling interpretation: sector-level revenue expansion appears to track expansion in operating capacity rather than reflecting costless digital scaling or operating leverage. This result directly addresses the capacity scaling gap highlighted in the Introduction by showing that CeFi growth appears to require parallel expansion in operating capacity rather than reflecting costless digital scaling.

The near-unit elasticity (β ≈ 0.97) indicates proportional scaling rather than increasing returns. This finding is inconsistent with narratives of costless digital scaling and instead suggests that CeFi revenue expansion requires commensurate growth in compliance, service, custody and infrastructure capacity. Recent evidence in banking digitalization similarly documents that profitability gains from digital transformation are conditional on the parallel expansion of operating capability rather than purely technological efficiency gains [8,13]. The result therefore positions CeFi growth within a capacity scaling framework rather than an operating leverage framework.

H4 is supported by Model M4, with an important qualification. Lagged SG&A growth (sga_yoy_lag1) is negatively associated with subsequent revenue growth (β = −0.397, p < 0.05), based on N = 4 observations after lagging. Notably, this negative lag association coexists with the positive contemporaneous association between SG&A growth and revenue growth in M1. Read jointly, the results are consistent with an intertemporal trade-off: SG&A expansion may coincide with growth in the same period, yet unusually rapid expansion may be followed by weaker growth in the subsequent period.

The coexistence of a positive contemporaneous association (M1) and a negative lagged association (M4) suggests that SG&A expansion may function as a growth accelerator in the short term but exhibits diminishing or corrective dynamics over time. This pattern is consistent with marketing overspending and metric unreliability arguments, where aggressive spending may temporarily inflate growth metrics before efficiency adjustments occur [29,45]. Given the aggregated design, this should not be interpreted as firm-level overspending behavior but as a sector-level adjustment pattern in which commercialization momentum and growth normalization co-evolve.

Model M4 provides suggestive evidence of a negative lag association between SG&A growth and subsequent revenue growth; however, the reduced number of observations after lagging (N = 4) precludes strong statistical interpretation. Observed relationships between SG&A dynamics and revenue measures may be bidirectional, since revenue expansion itself can induce SG&A adjustments through scaling and budgeting mechanisms. Given the short time series structure, coefficient estimates and goodness-of-fit measures should be interpreted cautiously. The results describe statistical regularities within the observed sector trajectory and do not imply parameter stability beyond the sample. Models M1–M4 are not competing specifications but complementary representations addressing distinct analytical questions (growth, profitability, scale, and intertemporal dynamics). Model M4 is therefore not interpreted as a superior or final model, but as an auxiliary specification examining lag structure effects that are not captured in contemporaneous models.

While the regression models establish the empirical association patterns summarized in Table 4, the FCM layer translates these relationships into a structured scenario framework addressing the decision-support gap identified earlier.

The FCM scenarios are broadly consistent with the regression sign structure in the sense that expanding Operating Expenses Total and SG&A growth increase the simulated outcomes (revenue growth, Net Margin from Totals and Total Revenues), while negative shocks produce systematic contraction. This should be framed as scenario consistency rather than an independent hypothesis test because the FCM is parameterized from empirical association patterns. In addition, the simulated net margin response reflects the assumed network pathways (including efficiency- and scale-related nodes) rather than a direct claim that higher SG&A shares improve profitability. This distinction is important given the negative association between SG&A intensity and net margin in M2.

Importantly, the simulated amplification of positive and negative shocks reflects the interdependence of commercialization, capacity and margin structures in centralized finance. Because cost bases and revenue generation mechanisms are tightly coupled through regulatory execution and service reliability, reductions in operating capacity or commercialization momentum can propagate quickly through growth and margin pathways. The FCM representation makes this structural coupling explicit, illustrating how CeFi performance may exhibit nonlinear sensitivity to coordinated investment and retrenchment decisions.

Taken together, the findings extend the marketing accountability and digital intermediation literature by showing that SG&A-based commercialization indicators encode multiple analytically distinct performance dimensions—conversion efficiency, cost burden, capacity co-scaling and intertemporal adjustment. Treating SG&A as a single “marketing spending” proxy therefore obscures the differentiated mechanisms through which commercialization relates to growth, profitability and operating scale in centralized financial intermediation. Commercialization efficiency (revenue/SG&A) is positively associated with revenue growth, while commercialization intensity is negatively associated with net profit margin. Operating expense scale strongly co-scales with revenue scale, demonstrating structural interdependence rather than costless scalability. The link between revenue growth and SG&A growth is positive contemporaneously but negative thereafter, suggesting an intertemporal trade-off.

6. Conclusions

This study investigated the correlation between commercialization-related expenses, efficiency, and financial outcomes inside a centralized financial system, employing a combination of OLS regression models (M1–M4) and a fuzzy cognitive map (FCM) scenario layer. Despite being limited by a brief aggregated time series (2020–2025), the findings present a coherent and internally consistent pattern that elucidates the interaction between SG&A-related dynamics, revenue growth, profitability, and size.

6.1. Theoretical Implications

Beyond documenting statistical associations, the study contributes conceptually by demonstrating that SG&A-based commercialization measures do not represent a single latent construct but encode multiple analytically separable dimensions (efficiency, intensity, expansion dynamics, and capacity scaling). This differentiation clarifies inconsistent findings in prior research that treat SG&A either as a marketing proxy or as a generic cost category. By isolating these dimensions within a centralized financial system context, the study refines how accountability constructs should be interpreted under accounting aggregation constraints.

The findings enhance the literature on marketing accountability and SG&A proxies by clarifying the conceptual differentiation between commercialization intensity and commercialization efficiency. Efficiency-based metrics (e.g., revenue per unit of SG&A) are more indicative of revenue growth than absolute expenditure levels, providing empirical validation for the assertion that output conversion measures more accurately reflect the economic significance of marketing and commercialization efforts [10,11]. Furthermore, the findings indicate that SG&A should not be seen as a uniform marketing input. The inverse relationship between SG&A intensity and net margin (M2) underscores the profitability issues associated with cost burdens when commercialization expenditures outpace revenue growth. This corresponds with contemporary theoretical viewpoints that highlight SG&A as a collection of many functions—marketing, administration, compliance, and centralized finance—whose financial consequences may vary significantly [12]. The log–log elasticity findings (M3) further contest the assumptions of costless scaling or operational leverage at the sector level. Conversely, revenue growth seems closely linked to the expansion of operating capacity, indicating co-scaling rather than growth driven by efficiency. The differing immediate and delayed SG&A effects (M1 vs. M4) indicate an intertemporal trade-off: times of swift expansion driven by SG&A growth may be succeeded by phases of slower development, aligning with adjustment frictions and timing effects in aggregated financial systems. CeFi firms and regulators should assess commercialization expenditures by efficiency (output per unit of input) and intensity (cost burden relative to revenue) for policy and governance. Efficiency-based measures should guide performance and budgeting decisions. Additionally, capacity development strategies must account for the fact that operational expenses rise proportionally with revenue, requiring integrated growth and cost planning.

6.2. Practical Implications

From a managerial and policy standpoint, the findings advise against exclusively depending on SG&A growth or intensity targets as measures of commercialization success. Decision-makers in centralized financial systems must differentiate between increased expenditure and enhanced expenditure efficiency, as the latter correlates more robustly with revenue growth while mitigating negative impacts on margins. The research indicates that swift development of SG&A may yield short-term growth advantages, while possibly limiting future performance. Managers should thus augment growth-oriented commercialization initiatives with monitoring systems that evaluate delayed performance impacts and adjustment expenses. Efficiency ratios, as opposed to mere expenditure indicators, may yield more valuable insights for budgeting, performance assessment, and strategic planning. The significant correlation between operating expenses and revenues at a systemic or policy level suggests that growth initiatives will necessitate corresponding capacity investments. Anticipating automatic operating leverage or margin enhancement just from scale may be impractical in centralized frameworks where administrative and compliance expenses increase in tandem with activity levels.

6.3. Limitations and Opportunities for Future Research

A primary limitation of the study is the small sample size derived from the brief aggregated annual time series (2020–2025), which diminishes statistical power and constrains the reliability of traditional significance testing; consequently, the findings should be regarded as indicative patterns rather than statistically generalizable results. Numerous constraints affect the interpretation of the results. The principal limitation of the study is the short aggregated annual series, which limits inferential reliability and renders conventional significance testing sensitive to small-sample effects. The findings should therefore be interpreted as pattern-consistent rather than statistically generalizable. The limited sample size and aggregated annual design hinder statistical link inference and reduce statistical power, especially in lagged specifications (M4). Secondly, SG&A functions merely as a surrogate for commercialization efforts and may incorporate significant construct noise owing to its diverse components. The models rely on observable accounting aggregates only. Organizational, contextual, and behavioral factors not captured in financial statements may also influence the examined outcomes. This constraint is inherent to aggregated sector-level designs and limits explanatory completeness.

Methodological and contextual limitations limit this study. The short aggregated yearly time series (2020–2025) limits statistical power, causal inference, and pattern-consistent conclusions. Accounting-based proxies like SG&A may increase measurement noise and prevent functional cost decomposition. The analysis’s centralized financial system architecture may limit its applicability to decentralized or firm-level situations. These limits should be addressed while interpreting results.

The FCM study should be seen as a scenario-consistency exercise rather than an autonomous evaluation of statistical link mechanisms. Subsequent studies ought to corroborate these findings by the utilization of firm-level panel data, extended timeframes, and diverse marketing and commercialization proxies that provide a more distinct differentiation of functional cost elements [56]. These approaches would provide more rigorous statistical link testing, enhanced dynamic analysis, and greater understanding of how centralized financial arrangements influence the relationship between commercialization investments and financial performance [57]. Future research using firm-level panel datasets and longer horizons would enable more robust statistical inference and causal identification beyond the scope of the present aggregated annual design.

Implementation difficulties should be considered when interpreting policy and governance suggestions. Centralized financial institutions may struggle to adopt efficiency-oriented supervision frameworks due to organizational rigidities, legal constraints, information asymmetries, and structural limits. The institutional and behavioral constraints of the study may limit the applicability of the given recommendations.