Poster
Client Context
BrandSafway is a privately-owned construction and civil engineering company, headquartered in Kennesaw, GA. Aluma Systems, a subsidiary of BrandSafway, rents forming and shoring equipment for various projects such as hotels, high-rise buildings, stadiums, and airports in the United States and Canada. Aluma holds $530 million of inventory across 13 branches in North America with 5,000+ SKUs. When an order is placed, if a branch does not have all of the equipment needed for a project, Aluma must pay the transportation cost to ship the missing items from the nearest available branch in its region using a third-party carrier. These shipments are called branch-to-branch shipments.
Project Objective
In the past four years, Aluma has experienced a steady increase in their annual logistics costs. The motivation behind this project is to reduce branch-to-branch shipments, which is the leading contributor to these costs. These shipments occur when items are not located in the region where they are later demanded.
The overall objective is to determine a way to decrease the branch-to-branch shipments. One opportunity to achieve this was to analyze the inventory mix at each branch and propose new, data-driven inventory level reference points for each item that each branch can maintain.
An additional opportunity was to analyze the utilizations of individual item-branch combinations (IBCs) to identify further cost saving opportunities.
- The purpose of the first additional analysis is to identify underutilized items and offer inventory quantity changes that can result in the most cost savings. For these IBCs, there is a cost tradeoff between increasing free cash-flow by selling off excess inventory and incurring shipping costs when needed.
- The second analysis identifies IBCs with a utilization greater than 1 and generates the cost of acquiring enough new inventory to meet demand over certain periods of the year. A utilization greater than 1 indicates that a branch-to-branch shipment was required.
Design Strategy
The design strategy consists of four stages: a cost regression model, an optimization model, random demand generation, and a simulation.
Our initial solution to decrease branch-to-branch spending is a recommended inventory mix, which is a list of inventory levels for each item at each branch and regional level, generated by an optimization model. This optimization model computes the minimum branch-to-branch shipping cost the company could have achieved if they knew all demand in advance by deciding where inventory should have been initially allocated between branches and how it should have been transferred throughout the year. The model incorporates the results of a regression analysis on the cost of branch-to-branch shipments.
Using a python simulation and probabilistic future demand scenarios, the robustness and accuracy of the inventory mix is determined by analyzing the variability of the branch-to-branch costs from the simulations. The simulation acts a real-world simulation to find a realistic branch-to-branch cost savings for the future. The simulation was performed on demand variations as well as a pessimistic 20% decrease and an optimistic 20% increase in historical demand.
The additional objective, a utilization analysis, uses a python model to calculate the average annual utilization. The first analysis determines inventory changes to increase utilization to certain thresholds and for each of these inventory change values, calculates the potential cost tradeoff using the allocation costs and shipping costs for each IBC. The second analysis determines the required inventory changes to reach a fill rate of 95% and the cost of adding that inventory.
Deliverables
The final deliverable from the optimization model is a suggested inventory mix for each item type for each branch and region. These mixes will serve as reference points to maintain going forward. The deliverable is in the form of a Power BI report.
The second deliverable is a list of underutilized IBCs with inventory changes that result in the best cost savings. The third deliverable is the required inventory changes to reach a 95% and 99% fill rate and the total projected cost of the added inventory.
The optimization model and simulation also act as deliverables as they allow for what-if scenarios to be analyzed. Such scenarios include where inventory levels and/or demand are changed. The models can then predict how shifting, acquiring, or retiring inventory affects the operations of the system and estimate the costs incurred. They can also evaluate pessimistic and optimistic demand scenarios (i.e., percentage-based increases or decreases in historical demand values). If a large swing in demand is expected in either direction, the models can estimate how the system will perform as a result.
Value and Impact
The optimization model provides value by generating robust inventory mixes that Aluma can work towards implementing and can be used as data driven reference points to maintain to decrease branch-to-branch shipments in the future.
The optimization model's optimal allocation results in an expected cost a ¼ of the determined historical branch-to-branch cost. The simulations generated a low standard deviation for the branch-to-branch costs which indicate that the suggested mix is robust to demand fluctuations. This result, along with the negligible increase in costs associated with a 20% increase in demand, indicate that overall inventory levels are not a limiting factor and inadequate inventory mixes are the leading cause of branch-to-branch shipments. The optimization model also provides value since it can be used as an interactive, adaptable inventory planning tool.
The Utilization Additional Analysis found that making the inventory changes necessary to apply a 5%, 15%, or 25% increase to utilization (whichever would yield the largest savings) to each of the IBCs with low utilization would eliminate excess inventory with a total allocation cost of $72M. This is inventory that can be sold or not replaced after retirement with minimal effect on branch-to-branch costs (as these items are unlikely to be shipped) while still holding enough inventory to meet future demand.
The Overutilized Inventory Analysis is valuable as it shows how much inventory needs to be added at each branch to meet a 95% or 99% fill rate to minimize branch-to-branch shipments. This analysis found a purchasing cost of $6M to meet the 95% fill rate.
