Normally transitions between different VDM dialects go from VDM-SL towards VDM++ or VDM-RT. In this paper we would like to demonstrate that it actually can make sense to move in the opposite direction. We present a case study where a requirement change late in the project deemed the need for distribution and concurrency aspects unnecessary. Consequently, the developed VDM-RT model was transformed to VDM++ and later to VDM-SL. The advantage of this transformation is to reduce complexity and prepare the model for a combined commercial and research setting.