It is generally known that software system development lifecycle (SSDL) should be managed adequately. The global economy crisis and subsequent depression have taught us certain lessons on the subject, which is so vital for digital transformation, for Industry 4.0. The paper presents the adaptive methodology of enterprise SSDL, which allows to avoid "local crises" while producing large-scale software. The methodology is based on extracting common ERP module level patterns and applying them to series of heterogeneous implementations. The approach includes a lifecycle model, which extends conventional spiral model by formal data representation/management models and DSL-based "low-level" CASE tools supporting the formalisms. The methodology has been successfully implemented as a series of portal-based ERP systems in ITERA oil-and-gas corporation, and in a number of trading/banking enterprise smart applications for other enterprises. Semantic network-based air traffic planning system, and a 6D-model-driven nuclear power plant construction support system are currently in progress.