Dealing with Last-Mile Analytics

Evidence from Indonesian Tax Administration through Practice Research

Abstract

This study employs a “practice-as-research” approach to investigate the role of analytics as a pivotal component of Indonesia’s tax administration system, specifically focusing on addressing tax gaps. Tax analytics systems and applications operate centrally to generate data that support several tax administration functions, including taxpayer registration, compliance monitoring, dispute resolution, and law enforcement. However, the tax officers—who serve as end-users of this data—frequently encounter the “last-mile problem”, where the data provided is not immediately actionable. Consequently, such tax officers are often required to develop their own data pipelines to further process and analyze the data before it can be effectively used for decision-making. This study identifies two categories of last-mile issues: those that can be eliminated and those that can only be mitigated to a limited extent. Two key recommendations are proposed to address these challenges. First, existing analytics applications should enhance taxpayer profile data by integrating the most comprehensive analytics outcomes, including compliance risk profiling. This can be achieved by implementing a “reverse-ETL” approach to improve existing analytics applications, facilitating the seamless flow of processed data back into operational system data. Second, the study advocates for more flexible self-service analytics platform for scenarios where last-mile challenges are unavoidable. This could be an analytics sandbox or a data-as-a-product approach that leverages containerization to enable tax officers to process and analyze data independently. These recommendations aim to improve the efficiency and effectiveness of Indonesia’s tax administration system by addressing the critical last-mile challenges faced by tax officers, thereby enhancing the overall utility of analytics in supporting tax-related decision-making processes.