The first part of the article provides a picture of the business case of demand forecasting and how it concretely affects our supply chain in a positive way. The second part dives deeper into how to generate accurate forecasts from a data analytical perspective.

Setting the Context: The Business Model for Demand Forecasting

Within an organization it is important to drive projects based on the underlying business model. So it is ensured that innovation is enabled where it is needed. To do so, we use the magic triangle of business models as a variation of The St. Gallen Business Model Navigator to describe and balance the business constraints and ensure a high-quality outcome of the project. Figure 1 shows the magic triangle for our demand forecasting use case. If you are interested in the magic triangle for business models in general and how to approach data-driven decision making projects from a holistic view, have a look at my previous article.

The Business Model for Demand Forecasting in Logistics

Who is our Client?

The first dimension is the clients within the company for our demand forecasting business model. Multiple stakeholders might be interested in the use case to drive their goals and objectives. The logistics manager is our main stakeholder since we want to use demand forecasting to optimise the supply chain. There might be other stakeholders involved as well such as the sustainability officer who wants to improve the company’s sustainability. The other three dimensions of the magical triangle help us to get a clear picture of the fundamental value aspects of demand forecasting. The goal of supply chain management is to ensure the smooth flow of goods, information and finances across the supply chain to meet customer demands while optimising costs and maximising overall efficiency. To achieve those goals, the following three competitive objectives for logistics must be balanced.

• What is the effect on the customer? Customer intimacy focuses on building strong and deep relationships with customers. It involves understanding customers' needs, preferences and behaviours to tailor products, services and interactions to meet their requirements. The value proposition is to guarantee those goals. We use the customer service level as a metric to measure customer intimacy.

• Which factors drive the value? Total Coverage Time is the time between an inventory item is first made available for sale and when it is replenished. By providing a more accurate demand forecast, the TCT can be reduced, which either results in a positive impact on multiple aspects of our supply chain costs or results in a higher customer service level.

• What is the effect on the financial performance? Supply Chain costs refer to the expenses incurred throughout the process of moving products or services from the initial production or procurement stage to the final delivery to the customer. It encompasses various elements involved in the supply chain - including procurement, manufacturing, transportation, warehousing, distribution and inventory management. The next section gives us a closer look at how cost structures and revenue streams contribute to the financial performance of demand forecasting.

Data Value Story: Demand Forecasting to Optimise Supply Chain Management (SCM) in Logistics

The Data Value Story for Demand Forecasting in a Nutshell

Demand forecasting is an important factor contributing to the mentioned objectives to ensure efficient SCM. With the use of statistical methods and data science, data-driven decision making (DDDM) brings many benefits to achieve a high-quality forecast such as:

Objectivity: Decisions are fact-based (data-based) and therefore reduce the influence of personal bias.

Improved efficiency: Data-driven decision making can help organisations make decisions faster and with less effort through automation.

Improved accuracy: Data-driven demand forecasting can extract statistical temporal relationships in the data, combine a large amount of information and is able to handle individual items flexibly.

Scalability: Data-driven decision making allows organisations to process large amounts of data and make decisions at scale for many items.

Assessing the Return on Investment (ROI) of a use case is crucial from an economic perspective. The aspects to consider must be aligned with our supply chain goals and objectives to ensure they are specific to our business. Translated to demand forecasting, we want to assess what is the value added by improving the forecast performance. The following graphic shows a selection of aspects in supply chain management that are affected and for which individual analysis and simulations should be assessed:

Effects of Demand Forecasting on Supply Chains and Related Domains in Manufacturing and Retail

• Raw Materials and Procurement: Improved demand forecasting can lead to more efficient procurement strategies by reducing over-purchasing or emergency buying and taking advantage of bulk discounts or favourable market conditions.

• Inventory Holding Costs: With accurate demand forecasting, you can minimize the inventory levels required to meet customer demand, thus reducing storage space, handling and capital costs.

• Production Overhead: When demand is accurately forecasted, production schedules can be optimized to reduce changeover times and improve machine utilisation, which in turn reduces overhead costs.

• Transportation and Logistics: Understanding demand patterns allows better transportation planning, potentially consolidating shipments and optimizing routes, which can reduce transportation costs.

• Lead Time and Responsiveness: Accurate demand forecasting enhances the ability to respond quickly to changes in demand, which can reduce lead times, lower expedited shipping costs and improve customer satisfaction.

Based on the complexity and maturity of your business, different methods for ROI assessment are appropriate. Common ones are:

• Using analysis and simulations specific to your business and supply chain

• Using the "business value / effort matrix"

• By general rules, e.g., findings by Gartner (see below)

For the sake of this article, we will keep it simple by using a more general rule, based on the findings provided by Gartner. Gartner reports that for every 1% forecast improvement, a consumer goods company could achieve a 2.7% reduction in finished goods inventory (days), a 3.2% reduction in transportation costs, and a 3.9% reduction in inventory obsolescence. In the example of a retailer, we assess the ROI based on the impact of improvement by demand forecasting on their supply chain. We apply those general rules to our specific costs for the single aspects (e.g., finished goods inventory costs: 7 days avg.; transportation costs: 50 Mio €; inventory obsolescence costs: 17 Mio €) of our supply chain process. By doing so, the following figure shows the related impact of a 1% improvement in demand forecasting accuracy.

Impact of each 1%-Improvement of Demand Forecasting Accuracy

With a clear picture in mind about the business aspects of demand forecasting in logistics, we want to dive deeper in the next step how to achieve a more accurate forecast from a data analytical perspective. If you are interested in the general process of the data science lifecycle and how to define the right data science architecture for your project, have a look at my previous article.

Holistic View for Data-Driven Decision Making Projects

Summary

Demand forecasting stands as a linchpin in supply chain management, shaping efficient logistics and data-driven decision making. Leveraging modern forecasting techniques not only refines our understanding of customers but also paves the way for optimized operations and financial benefits. As demonstrated, even a modest improvement in forecasting can yield tangible results by highlighting the profound impact of embracing data science in logistics. As we advance, the precision in forecasting will inevitably shape the future of supply chain optimization.

We Support Your Data-Driven Journey

This is the third of an article series on data-driven decision making, data science, AI and machine learning. In the previous article, we learned how to integrally approach any data-driven decision making project. In the next one, we will take a closer look at how to create accurate forecasts in sales and demand forecasting.

We would be pleased to discuss your interest in these topics with you in person. Contact our Munich office or write an email to David Tiefenthaler.

foryouandyourcustomers is happy to support your company in becoming more data-driven.