The goal of any Lean organization is to understand customer value and to focus its efforts to continuously increase it. Lean applies to every business and every process and so is applicable for New Product Development (NPD) in Automobile industry, where the major output is the vehicle part numbers. Part numbers are generated based on the variant tree finalized. Customer requirements, benchmarks and organization assets such as lessons learnt and historical information provide input to the variant tree. Parts numbers for a particular model are generated during the concept Bill of Materials (BOM) stage and after which it exists during the complete product life cycle. Part number generation includes considerable effort by the design team, the validation team, and also includes overheads on the Product Life cycle Management (PLM) system.This paper focuses on minimizing the generation of part numbers for a particular model based on past one year production data of the base model or an equivalent model. Rank Order Clustering (ROC) technique applied to Group Technology is being used as a heuristic approach to identify the group of aggregate specifications which has been part of the volume produced and which has not. Set difference is further used to shortlist the aggregates for elimination. Then this array is further reinforced by verification against the Production volume. Flow of results of the ROC method based on past data to the product under development is also discussed.