A contingent value of bricolage strategy on SMEs’ organizational resilience: lessons from the COVID-19 pandemic

The COVID-19 pandemic created severe supply and demand shocks and operational constraints that disproportionately affected SMEs due to resource scarcity, limited crisis experience, and constrained access to finance. Organizational resilience (OR)—the capacity to absorb shocks and adapt while maintaining core functionality—is critical for SME survival and continued innovation during crises. Prior crisis management research highlights entrepreneurial action, flexible decision-making, contingency planning, communication, and collaboration, but analyzing isolated actions may miss the multifaceted strategies SMEs need under constraints. The study proposes that bricolage—improvisational, flexible recombination of available internal and external resources—offers a holistic, adaptive crisis-response approach that can enhance SMEs’ OR. The research question asks: Under what conditions does a bricolage strategy improve SMEs’ organizational resilience in crises? The study posits that the effectiveness of bricolage for resilience is contingent on external institutional measures (government financial and operational support) and internal strategic actions (business process innovation). Using a large-scale survey of Korean service-sector SMEs during COVID-19, the study aims to empirically assess these contingencies and inform crisis management theory and practice.

Theory and hypotheses integrate resilience and bricolage literatures. OR is defined as a firm’s ability to positively adjust and maintain functioning before, during, and after adversities through absorptive and adaptive capacities. Resilient firms exhibit adaptive capacity, flexibility, improvisation, and strategic resource reserves. Innovation continuity is central to long-term success; discontinuing innovation during crises erodes performance by losing knowledge assets. Bricolage (Lévi-Strauss; Baker and Nelson) entails improvisation, using resources at hand, and creative recombination. It enables SMEs to cope with resource penury by repurposing underutilized internal resources, acquiring low-cost external resources, and leveraging relationships. Prior measures (e.g., Senyard et al.’s bricolage scale) operationalize these behaviors, though context-specific refinements are encouraged. The study argues bricolage should bolster OR in innovation activities during exogenous shocks (H1). Government support is a vital policy lever during crises; financial (subsidies, loans, guarantees) and operational (tax relief, HR/training, infrastructure, regulatory easing, market/export support) measures can alleviate constraints and foster recovery, potentially amplifying bricolage’s effect on OR (H2). Business process innovation—significant changes in operational processes across functions—can enable strategic pivots, but for resource-constrained SMEs it may also strain limited resources, potentially dampening the resilience gains from bricolage (H3).

Design: Quantitative cross-sectional analysis using the Korean Innovation Survey (KIS) 2021, aligned with Oslo Manual 2018 and Community Innovation Survey guidelines. Informants were CEOs/top managers. Population: South Korean service-sector firms in 2020. Sample: Initial 4000 firms across 33 service industries (from 60,111 in Statistical Business Register). Inclusion criteria: established before 2018, >10 full-time employees, ≥3 years in operation by 2020. Stratified, systematic sampling by STEPI. Exclusions: industries with no variation in the dependent variable (OR) to ensure logistic regression estimability; firms in unclassified industry to avoid collinearity. Final sample n=3179 SMEs with complete construct data. Descriptives highlight diverse service industries and firm sizes (e.g., most 11–50 employees). Measures:

• Organizational Resilience (DV, binary): OR in innovation activities operationalized via four items capturing absorption and adaptation during COVID-19: planned innovation activities dropped, deferred, implemented as planned (absorption), or newly added (adaptation). OR=1 if implemented as planned or newly added; OR=0 if dropped or deferred.
• Bricolage strategy (IV): Sum count of managerial actions taken using resources at hand during COVID-19 across 12 possible responses (e.g., remote work, new products/services, changes in production methods, intra-/inter-firm virtual networks, electronic transactions and non-face-to-face marketing, etc., plus other actions). Higher counts indicate greater bricolage engagement.
• Government support (moderators, continuous): Respondents rated the influence (1–5 Likert) of specific support types. Aggregated into two constructs: financial support (direct subsidies; finance support via loans/guarantees/investment/insurance) and operational support (tax measures, contact-free infrastructure, HR/training, regulatory easing, market development, export assistance). Each construct measured as the average influence across its items.
• Business process innovation (moderator, count): Number of distinct process innovations initiated over the past year across six functions (manufacturing/engineering/testing; distribution/logistics; marketing/sales; ICT systems; administration/management including HR, finance, procurement, SCM; product/service development processes). Controls: Firm size (employees, 2020), export intensity (exports/sales, 2020), listed firm status, clustering firm status, new venture certification, and industry dummies (two-digit KSIC). Analytical approach: Logistic regression with odds ratios reported. Models: controls (Model 1); main effect of bricolage (Model 2); interactions with government financial (Model 3) and operational support (Model 4); interaction with business process innovation (Model 5); full model (Model 6). Multicollinearity checks via VIF before and after mean-centering variables in interaction terms (no VIF>10). Maximum likelihood estimation; interaction effects visualized via predicted probabilities.

• H1 supported: Bricolage strategy positively associated with OR. Model 2 odds ratio (OR) = 1.761, p<0.01.
• Government financial support (GFS): Main effect positive on OR (Model 3 OR=1.797, p<0.01). Interaction with bricolage negative and significant (Model 3 OR=0.592, p<0.01), indicating substitution—financial support diminishes the resilience gains from bricolage. Predicted probabilities: at low GFS (mean−1 SD), increasing bricolage raises OR probability from 19.2% to 51.2%; at high GFS (mean+1 SD), it increases only from 21.6% to 29.6%.
• Government operational support (GOS): Main effect positive on OR (Model 4 OR=3.130, p<0.01). Interaction with bricolage not significant (OR=0.666, n.s.). Overall, no support for H2’s positive moderation across support types.
• Business process innovation (BPI): Strong positive main effect on OR (Model 5 OR≈124.640, p<0.01). Interaction with bricolage negative (Model 5 OR=0.545, p<0.10), supporting H3 that BPI attenuates bricolage’s positive effect on resilience. Predicted probabilities: at low BPI (mean−1 SD), increasing bricolage raises OR probability from 0.7% to 15%; at high BPI (mean+1 SD), increasing bricolage slightly reduces OR probability from 85% to 82.7%.
• Controls: Firm size positively related to OR (OR≈1.001 per employee, p<0.01). New venture status positive (Model 1 OR=2.501, p<0.01), suggesting greater entrepreneurial resilience. Listed firm positive at marginal significance (p<0.1 in Model 1). Industry effects controlled.
• Model fit: Chi-squared improves across models (e.g., 382.86 in Model 1 to 2114.96 in Model 5) with corresponding log-likelihood gains, indicating strong explanatory power when including moderators.

Findings confirm that bricolage—improvisational recombination and use of available resources—enhances SMEs’ organizational resilience in innovation during crises, aligning with theories linking resilience to improvisation, flexibility, and adaptive capacity. However, the resilience benefits of bricolage depend on contextual factors. Government financial support substitutes for bricoleurs’ self-sustaining resource orchestration, thereby diluting bricolage’s incremental contribution to OR, whereas operational support improves OR without altering bricolage’s effect. Concurrent pursuit of business process innovation appears resource-intensive for constrained SMEs and can divert attention and resources from improvisational recovery, attenuating bricolage’s resilience benefits even though BPI, by itself, is strongly associated with higher OR. These results clarify when bricolage is most effective: under limited financial aid and manageable levels of process change, where improvisation can translate into resilient innovation continuity. The distinction between externally reactive OR (absorptive/adaptive responses to shocks) and internally proactive BPI (structural process changes) underscores the need to balance short-term survival-oriented improvisation with longer-term process transformations. The study broadens crisis management understanding by integrating firm-level strategies with institutional interventions and demonstrating their interactive impacts across diverse service sectors.

The study shows that SMEs employing bricolage during COVID-19 achieved higher organizational resilience in innovation, but these gains are contingent. Government financial support reduces the marginal resilience benefits of bricolage, government operational support boosts resilience without moderating bricolage’s effect, and business process innovation—while strongly beneficial overall—weakens bricolage’s positive impact when pursued simultaneously. Contributions include: empirical validation of the bricolage–resilience link; identification of contingent roles of policy supports and process innovation; and a holistic framework integrating managerial and institutional responses. Practically, SMEs should prioritize bricolage for short-term resilience, leverage operational supports, apply financial aid judiciously to avoid dependence or administrative burdens, and phase resource-intensive process innovations to avoid undermining improvisational recovery. Future research should address causality with longitudinal/experimental designs, explore nonlinearity and thresholds (e.g., excessive bricolage), and adopt process-based and configurational views of OR across pre-, during-, and post-adversity phases.

• Cross-sectional design and potential endogeneity limit causal inference between strategies and resilience outcomes.
• Assumed linear relationships may overlook nonlinear or threshold effects (e.g., excessive bricolage causing resource dispersion and scalability issues).
• Measurement context is crisis-specific (COVID-19, Korean service SMEs), which may affect generalizability across institutional environments.
• Resource-intensity and timing of business process innovation not directly modeled; dynamic sequencing effects remain unexplored.
• Reliance on self-reported measures may introduce common-method bias despite model controls.