New tools make multi-brand, multi-node distribution manageable

  • August 30, 2022
Abstract connected shapes

For multi-node distribution networks to deliver merged, multiple-brand shipments reliably to retail customers, you must manage complexity at a world-class level. Fortunately, new tools make this demanding process easier.

The benefits of combined multi-branded product distribution are self-evident. Making it happen — not so much.

When your organization houses several well-known brands, the temptation is to allow each brand to operate autonomously — receiving, processing and fulfilling its own orders. This arrangement may make serving each brand’s customers more manageable, but it inevitably leads to inherent inefficiencies and wasted resources.

Developing a new distribution service network allows delivery of products — from multiple brand divisions — to customers in full truck shipment volumes. Providing a smooth transition demands the development of new storage space management policies at each distribution center. Also, the unique variables each delivery node presents call for specific methods to accomplish the packaging and shipment of multi-brand loads to each customer.

A running start makes for quick gains

It can be challenging to consolidate brand products at distribution centers (DCs) when each brand unit has a wide range of product sizes. Consider the following two essential project objectives:

  1. Determine which products to store at each DC and the replenishment policies (cycle stock and safety stock) by SKU in support of sales at set service levels
  2. Develop fair-share space mix allotments by business units and product families

The Integrated Demand & Supply Planning (IDSP) team at NTT DATA suggests either a SQL/Excel-based flow model or a third-party application-based model. In many cases, the IDSP team and our clients choose a SQL/Excel model to quickly provide the general answers needed to speed up program launch. We use this model to set initial space boundaries for each brand in each DC. Organizations implementing both methods most often have a set of brands that have little to no experience with multi-node fulfillment. This first step is an ideal starting point for the process.

Quick gains are good, but not when greater efficiencies are available

With a project off the ground and the low-hanging efficiency fruits duly picked, the focus turns to a third-party application-based model. It provides a detailed simulation that can forecast requirements with much higher accuracy. This model includes additional parameters on the previously mentioned minimum order quantity (MOQ), lead times and other variables. The application-based model also provides mechanisms for generating multi-echelon inventory optimization (MEIO) scenarios.

An application-based model creates new opportunities

Application-based model approaches perform best-fit, probabilistic forecasting. Input is 36 months of order-line historical demand. Planning application capabilities allow for the number of parameters, evaluating the outcome of the scenarios modeled. Parameters usually factored into models are MOQ, lot size, demand variability, replenishment lead time and supply variability. The model then calculates the recommended inventory levels for products at different locations and at various levels of aggregation.

The Chainalytics IDSP team often develops application-based models in cooperation with ToolsGroup. We use the SO99+ service optimization platform because of its unique forecasting capabilities and high customer retention rate. The ToolsGroup model incorporates a proprietary approach that uses stochastic models. Minimal maintenance settings fine-tune best-fit forecasting techniques. These capabilities far exceed traditional forecasting methods.

Application-based models support probabilistic forecasting and service-level optimization capability. They support specific objectives, such as minimizing cost, space (volume) and pallet positions. Service levels in this approach can be set as fixed, bounded or left “free” to allow the optimizer to find the best portfolio solution. Usually, it’s one that yields the highest service levels with the lowest inventory investment.

More (and better) data and robust tools yield more precise results

When compared to a SQL/Excel-based model, an SO99+ model offers additional scenario comparison capabilities. They leverage MOQs, lead times, product costs, estimated delays, optimized order-up-to levels, reorder points, safety stocks and estimated space consumption. This tool is powerful because it can optimize service-based desired outcomes using MEIO scenarios.

For companies that want to operate multiple DCs, we usually recommend introducing an MEIO strategy. It’ll help you achieve targeted service levels while balancing inventory among DCs. This strategy enables the network to keep inventory levels low, move inventory quickly and support high service levels for customers.

Using a ToolsGroup model in a median multi-brand, multi-node scenario will show an average of 5% reduction in space versus SQL/Excel-based models, at a targeted 93% service level (regarding non-stock-out probability). In-depth studies of different MEIO scenarios with the application-based model can show potential savings of 10% in space and inventory costs.

With a more precise application-based model, our team can build and compare multiple inventory replenishment scenarios with a quicker turnaround. Doing so also provides a sophisticated data-driven solution. It enables clients to make better-informed decisions that consider multiple variables and time constraints.

— By Hugh Walters

Subscribe to our blog


Related Blog Posts