Start with the work that fails or gets delayed. Record how often it happens, who it affects, and whether the cause is configuration, the business process, or a limitation of the system.
Compare repair and replacement
Use the same criteria for both options: cost, time, operational risk, and support for essential tasks. Include migration, training, and any period when teams must work across both systems.
Define the threshold for replacement
Test whether a reasonable fix can eliminate a recurring problem, such as double entry on common transactions. If an underlying limitation prevents a safe fix, document it and make that scenario part of the vendor evaluation.
Illustrative photo by William Liu on Unsplash.