Axtria Insights

Axtria Insights

CASE STUDY
A US Leading Home Improvement and Constructions Products Retailer, Maximizes Share of SKU Distribution Through Automated Distribution Centers

Situation

  • A large retailer of  home improvement and construction products  imports about 5000 SKUs to be stocked for sale in over 2,000 stores across USA
  • SKUs arrive at one of four different ports and are transported from the ports to the stores.
  • The transport process routes SKUs to the stores either through ADCs (Automated Distribution Centers) or MDCs (Manual Distribution Centers).
  • ADCs are highly automated and do not stock SKUs. MDCs on the other hand stock SKUs and involve expensive manual handling for stocking and subsequent shipment of SKUs from MDC to stores.
  • Ideally, the client wanted to have most of the SKUs flow through the ADCs. However during the lead time for SKUs to be transported to the US, the demand for SKUs at store level can change and as a result, some SKUs need to be stocked at MDCs, for which retailer incurs a higher cost. There could be multiple factors that cause the MDC path to be utilized.

Challenges

  • The retailer wanted to understand and quantify the different root causes for the MDC path being utilized substantially more than the ADCs
  • The client also wanted a better understanding of the forecast variance drivers and ability to forecast the Port flow path allocation to the ADC and MDC

Approach

We followed a two-step process to meet the objectives of the project. 

Root Cause Analysis

Initial step for RCA was a discovery phase; during this phase relevant data required from retailer was identified and collected. Various possible reasons/factors were also discussed with the retailer which can impact the utilization of MDC path.  Various statistical models were tried based on the nature and availability of data, out of which Logistic regression and OLS regression models were performed.

Port path flow modeling and validation

Simulation technique was used to capture demand /supply flow at various stages i.e. stores, DCs and ports.

Result

Through the improved forecasting model the client was able to bring down the variance from planned flow model to 7% from 10%.

Maximize Share of Sku Distribution to Stores via Automated Distribution Centers