SME crisis management and performance: leveraging algorithm supported induction to unravel complexity

• Published in: Journal of computational social science Vol. 8; no. 3
• Main Authors: Graham, Byron, Matikonis, Karolis
• Format: Journal Article
• Language: English
• Published: Singapore Springer Nature Singapore 01.08.2025
• Subjects: Big Data/Analytics; Complex Systems; Computational Linguistics; Research Article; Simulation and Modeling; Social Media; Social Sciences; COVID-19; Crisis management; Algorithm supported induction; Machine learning; SME performance; Complexity theory; Configuration analysis
• ISSN: 2432-2717; 2432-2725; 2432-2725
• DOI: 10.1007/s42001-025-00383-x

This study contributes to the future directions of SME crisis management literature through algorithm supported induction by exploring the complex relationships between SMEs’ strategic responses to the COVID-19 pandemic and their performance. Using data from the UK Longitudinal Small Business Survey, decision tree algorithms and explainable artificial intelligence techniques reveal how configurations of strategic actions and contextual factors shape performance outcomes. The analysis also uncovers dominant determinants and highlights previously overlooked non-linear and asymmetric relationships. Key findings emphasise the critical roles of responses to lockdown measures, utilization of the furlough scheme, and the interplay of firm size and age, which interact in complex configurations exhibiting asymmetry and non-linearity. This understanding provides a basis for informing future research directions, hypotheses, and strategies for SMEs to navigate crises and enhance resilience.