Overcoming the limitations of social media analytics: Dynamic capabilities and workaround strategies for inclusive and resilient public service continuity

Contemporary disaster management calls for inclusive approaches to the design and use of analytics systems, particularly those built on publicly available data used to train artificial intelligence. While widely adopted, social media analytics often generalize community concerns, sentiments, and needs, overlooking key characteristics such as age, occupation, interests, and language. These systemic limitations are especially problematic during crises, when public organizations must implement community-based disaster risk management (DRM) strategies grounded in listening, understanding, and responding to diverse populations. However, overreliance on aggregated analytics can marginalize vulnerable groups and ignore critical nuances. Drawing on interviews with public servants and software developers who use social media as an information source, this study reveals how current analytics tools fail to automatically acknowledge and provide insights regarding vulnerable populations (e.g., linguistic minorities or younger members of society) voicing their concerns in social media conversations. Recognizing this limitation, public organizations employ adaptive, absorptive and innovative dynamic capabilities through workaround strategies to understand, engage, and serve marginalized populations. By doing so, public organizations ensure diverse voices are included, and services remain resilient during disruptions. Practically, the findings inform the design of more inclusive analytics tools and crisis communication strategies; theoretically, they extend the application of dynamic capabilities to public sector innovation in disaster contexts. This study offers new insights into the intersection of social media analytics and DRM practices, highlighting the critical role of dynamic capabilities in fostering inclusive, resilient public services throughout crisis cycles.