GitaCloud helped a leading global automotive company sense remanufacturing demand better by deploying advanced predictive statistics based on regression analysis between cars sold and cars coming back for repair / remanufacturing of key components. The customer can now ensure they continue to plan for the right levels of parts and remanufacturing capacity across the globe as failure rates evolve.

Business Context: 

  • Multiple car configurations have resulting different component configurations. These components have different failure rates in each region based on road quality, weather, driving habits, and other factors.
  • Customer has regional remanufacturing centers to convert faulty components received back to new component quality spec level. Depending on the failure reason, the remanufacturing operation may or may not require parts and needs varying amount of labor capacity.
  • Customer needs to accurately sense demand for parts and labor capacity required in their regional remanufacturing centers. In case of a demand – supply imbalance, they either have to 1. ship faulty components to another region with available capacity resulting into higher cost and customer satisfaction issues; or 2. they may have a situation where they have labor capacity surplus in a region resulting into wasted manpower.
  • To balance the remanufacturing demand with remanufacturing supply by region, Customer needs robust correlation between good cars shipped by car level configuration and broken parts being received later for those good cars. This correlation needs to form the basis for projecting future arrivals of broken parts based on car build plans by car model / configuration. 

Pain Points with existing solutions:

  • Customer was managing this process in an excel based model. It was a complex model with hundreds of columns across multiple worksheets to track region specific component specific data (given lack of multi-dimensional features in excel).
  • Customer also had to deal with data integrity issues as any calculation change would need to be replicated in dozens of places and any resulting errors are hard to troubleshoot later and result into over/under capacity situations.
  • The data required to comes from several sources (homegrown ERP, other excel models like S&OP, etc.) and it is extremely cumbersome to update the excel models on a weekly basis.
  • Customer did not have capacity to analyze failure rates by failure reasons as constantly evolving and maturing product engineering means certain failure codes which were prominent in the fast become less of a problem or are eliminated altogether resulting into a non-linear and hard to forecast returns rate for parts in future.
  • The future broken parts forecast also needs to drive Warranty Analytics and other downstream processes in planning. Excel based solution hampers end to end process integration.

How GitaCloud helped:

  • GitaCloud team engaged with the customer to understand business pain points, requirements, and current excel based demand sensing models. GitaCloud supply chain experts quickly decomposed the excel model and requirements down to a detailed blueprint with demand modeling improvement ideas and data/integration using a cloud platform.
  • GitaCloud data analysts and demand model builders led the model build and completed a complex model development process in an aggressive time frame.
  • GitaCloud set-up a flexible demand model to convert car level configurations to corresponding components using flexible Bill of Material BOM mapping logic. This allowed the team to convert car configuration level past shipments data to corresponding components data.
  • The outbound data is compared with incoming faulty components data to calculate historical return factors. Advanced Linear regression techniques were used to identify slope and intercept of historical return factor data set to enable forecasting of return factors in future. Future car level build plan data is used to then predict future faulty component arrival rates by region.
  • The arrival rates are compared to current faulty component inventory, actual components completion rates, work in progress inventory to project faulty component inventory levels in future. This enables the Remanufacturing planning team to flex labor capacity to ensure adequate incoming demand to keep labor utilized and at the same time ensure the projected inventory levels are manageable and service levels are maintained to ensure high customer satisfaction. 

Value Delivered: 

  • A cumbersome to maintain and complex to understand excel model has been transformed into a fully integrated and dynamic yet intuitive demand sensing model supported by a cloud based data visualization and interactive planning platform.
  • The resulting model has a set of dashboards and data loading views which enable planners and management to see historical trends, understand future, and run what-if simulations.
  • The customer team is pleased with the speed and quality of solution delivered.