Abstract

Small distributors often handle customer complaints through instant messaging, where complaint information is produced but not retained. Prior complaint-management systems are reported after the transition, for institutional users or structured datasets, without a measured account of the process they replace. This single-case study at a heavy equipment spare part distributor in Central Jakarta measures that pre-formalisation baseline and derives the system design and its evaluation indicators from it. Observation, WhatsApp history review and interviews with five staff and eight customers were triangulated into a case register with fixed counting rules: from January to March 2026 it held 70 complaints, of which 24 (34.3%) took more than three working days to resolve and 18 (25.7%) had histories administrators could not retrieve. A web-based system was developed through Rapid Application Development and tested with eight black box scenarios derived by equivalence partitioning. All scenarios produced the expected output, and each baseline problem has a corresponding capability at handover. These are functional and process-level results; no post-implementation data were collected and security, boundary, concurrency and volume testing were outside scope, so no improvement in service quality is claimed. The contribution is a measured baseline, a low-cost design traceable from each baseline problem to a tested function, and indicators with baseline values for later evaluation.