Bringing Visibility to a Complex Data Lake
Customers are demanding. They expect goods to arrive on time, at a minimum cost. Because retail margins are razor thin, keeping up with customer demands can be disastrous for profits if inefficiencies exist in your business. Maximizing delivery network efficiency and optimizing delivery schedules is a complex but tractable problem, elegantly handled by modern data science techniques.
This Consumer Retail equipment manufacturing company had an exceptionally complicated supply chain, with a distribution model spidered across an intertwined United States delivery chain. For the past fifty years their supply chain has evolved based on requests from sales managers to service customers from local facilities and requests from management to ship full trucks directly off the line. The result was a delivery network optimized to support each of these requests composed of facilities of varying size and operational capacity and trucks waiting to be filled on the line.
The data over fifty years of operation was messy and segregated, so the regional planners often relied on localized tribal knowledge to get customers product in time to meet SLAs. This intuition-driven delivery mechanism was ripe for system wide optimization to improve overall efficiency and reduce network costs.
Expero reduced freight and distribution costs over 12% by consolidating their supply chain from dozens of small, non-optimally placed distribution centers to an efficient six distribution facility network. In the new delivery network design, the company is able to ship goods to customers more cost effectively by guaranteeing a lower total freight cost. We then built a system to automatically adapt to changing customer demand in real time, to manage inventory across the new distribution network, pack freight trucks according to the current customer demand, and ship the goods with optimal delivery scheduling. All of this while meeting customer delivery SLAs.
Customers are demanding. They expect goods to arrive on time, at a minimum cost. Because retail margins are razor thin, keeping up with customer demands can be disastrous for profits if inefficiencies exist in your business. Maximizing delivery network efficiency and optimizing delivery schedules is a complex but tractable problem, elegantly handled by modern data science techniques.
This Consumer Retail equipment manufacturing company had an exceptionally complicated supply chain, with a distribution model spidered across an intertwined United States delivery chain. For the past fifty years their supply chain has evolved based on requests from sales managers to service customers from local facilities and requests from management to ship full trucks directly off the line. The result was a delivery network optimized to support each of these requests composed of facilities of varying size and operational capacity and trucks waiting to be filled on the line.
The data over fifty years of operation was messy and segregated, so the regional planners often relied on localized tribal knowledge to get customers product in time to meet SLAs. This intuition-driven delivery mechanism was ripe for system wide optimization to improve overall efficiency and reduce network costs.
Expero reduced freight and distribution costs over 12% by consolidating their supply chain from dozens of small, non-optimally placed distribution centers to an efficient six distribution facility network. In the new delivery network design, the company is able to ship goods to customers more cost effectively by guaranteeing a lower total freight cost. We then built a system to automatically adapt to changing customer demand in real time, to manage inventory across the new distribution network, pack freight trucks according to the current customer demand, and ship the goods with optimal delivery scheduling. All of this while meeting customer delivery SLAs.
Expero’s approach to this problem is multifaceted. We work extensively in supply chain optimization settings, and have a suite of tools to help us deliver with efficiency. This particular customer required a delivery network simulator capable of testing new distribution facility locations, and scoring them based on total forecasted cost to the company. The simulator operates on a customer demand pull-based methodology where the end consumers effectively draw inventory through the delivery network to their endpoints.
Having the simulator in hand, the next step was to use a geospatial optimization algorithm to place the future-state delivery facilities in the best possible (read: lowest cost) locations which still met customer delivery SLAs. Combining the geospatial optimizer with the forward delivery simulator, we built up an optimal delivery network. The deliverable output included reports to executive management on what the future state total cost would be, with anticipated operating expenses, and a total expected savings. It also included reports to senior management about where new facilities would be and how to assign customers to distribution facilities.
Now that the company has an optimal distribution network, the next most expensive operational characteristic is the physical distribution of goods over freight shipping modalities. Expero built a real time freight truck packing system which included truck routing and scheduling to optimize the use of the contract freight services the company utilizes. The deliverable output was a software system for supply chain planners and shipping managers to use which describes which items go on which trucks, and where each truck should make deliveries.
Expero’s approach to this problem is multifaceted. We work extensively in supply chain optimization settings, and have a suite of tools to help us deliver with efficiency. This particular customer required a delivery network simulator capable of testing new distribution facility locations, and scoring them based on total forecasted cost to the company. The simulator operates on a customer demand pull-based methodology where the end consumers effectively draw inventory through the delivery network to their endpoints.
Having the simulator in hand, the next step was to use a geospatial optimization algorithm to place the future-state delivery facilities in the best possible (read: lowest cost) locations which still met customer delivery SLAs. Combining the geospatial optimizer with the forward delivery simulator, we built up an optimal delivery network. The deliverable output included reports to executive management on what the future state total cost would be, with anticipated operating expenses, and a total expected savings. It also included reports to senior management about where new facilities would be and how to assign customers to distribution facilities.
Now that the company has an optimal distribution network, the next most expensive operational characteristic is the physical distribution of goods over freight shipping modalities. Expero built a real time freight truck packing system which included truck routing and scheduling to optimize the use of the contract freight services the company utilizes. The deliverable output was a software system for supply chain planners and shipping managers to use which describes which items go on which trucks, and where each truck should make deliveries.
