For demand sensing and forecasting in the supply chain, the ability to quickly ingest and analyze data, and subsequently make strong business decisions is crucial. While this is true across all aspects of supply chain management, it is especially important when tracking actual demand versus projected demand. This crucial need can be slowed down or impeded by issues such as a lack of end-to-end supply chain visibility, antiquated data management processes, or even inaccurate data.
Significant disruptions along the supply chain from external factors such as geopolitical events, supplier capacity issues, poor network inventory visibility, and constant changes in buyer behavior, make synchronizing demand and supply very difficult. This is further complicated by inaccurate data from dozens of disparate applications and enterprise systems within the organization, its partners, and its suppliers. Traditional forecasting methods struggle to keep up with rapid changes in global supply chains, often failing to predict demand accurately during volatile periods.
Companies have traditionally relied on historical data and internal systems for demand forecasting, but this approach is limited in its ability to respond to sudden market shifts. The ability to sense demand disruptions in real time and improve forecasting in this environment is difficult to achieve, especially if you want a high degree of customer satisfaction, and it also highlights the responsiveness needed to adapt quickly to unexpected changes. Companies that leverage demand sensing can emerge stronger and better positioned after disruptions.
An Introduction to Demand Sensing and Forecasting
Demand sensing and demand forecasting are both crucial aspects of optimizing supply chains, but they do have slightly different functions in their approach and focus. Demand sensing uses real-time data and analytics to identify and respond to immediate demand fluctuations, while demand forecasting uses historical data to predict future demand over a longer period (months or years). Different methods, such as statistical modeling and machine learning, are used to enhance the accuracy and adaptability of these processes. Both areas are crucial for companies when it comes to projecting sales, managing inventory, and coordinating replenishment. In the end, the goal is to accurately predict customer demand by using predictive models to forecast future demand.
From a metrics standpoint, companies need to accurately measure forecast versus actual sales, inventory turnover, stockout rates, inventory obsolescence, order fill rates, and on-time in-full percentage. When forecasting, it is important to predict demand for a particular product to avoid excess inventory and stockouts. Advanced analytics and AI tools provide granular insights into sales activities, inventory levels, and financial metrics, supporting more precise decision-making.
Recognizing the growing complexity of these demands, InterSystems surveyed 450 senior supply chain practitioners and stakeholders to examine key supply chain technology challenges, trends, and decision-making strategies across five key use cases: fulfillment optimization; demand sensing and forecasting; supply chain orchestration; production planning optimization; and environmental, social, and governance (ESG). This blog is Part 2 in our Optimizing Supply Chain Performance with Unified Data series, with a focus on demand sensing and forecasting.In the unified data survey, respondents were asked how they currently integrate and prepare disparate information for decision-making. Not surprisingly, 42% of respondents use manual methods, including spreadsheets, to integrate and prepare disparate information for decision-making. While spreadsheets can be incredibly useful and are clearly used by a lot of companies for planning purposes, they also have limitations.
As the picture above shows, spreadsheets are not a useful tool when it comes to decision intelligence. Decision intelligence is focused on improving decision-making by understanding how decisions are made and using AI and machine learning to optimize outcomes. In supply chain, an AI-enabled decision intelligence platform can optimally manage disruptions when and before they occur so companies can react faster and ensure that products are available when companies need them, while also monitoring engagement to improve sales outcomes.
Current State of Demand Sensing and Forecasting
One of the biggest issues with demand sensing and forecasting is that human intervention is often required. This is because AI often lacks the nuances to fully understand the complexity of demand patterns. So, while human intervention is required to bridge that gap, it can be both time-consuming and error-prone, especially if the data a company is relying on is bad. According to the survey results, when asked how they currently forecast demand, 36% of respondents indicated that they have several solutions that require staff input. Aside from the aforementioned issues with human input, the use of multiple systems often leads to disjointed, disparate data silos. When different systems are unable to communicate, decisions take longer to make and are usually not as accurate, leading to errors in demand sensing and forecasting. To maintain data accuracy and relevance, it is crucial that data is updated and transferred regularly.
The harsh reality is that the use of intelligent data platforms is not widespread. The survey revealed that only 27% of respondents have an intelligent data platform. This is most notable in logistics and transport (18%) and pharmaceuticals (19%) where less than one-fifth of companies are currently using an intelligent data platform. For these platforms to be effective, it is essential that all data is validated before being used in forecasting models to ensure consistency and accuracy.
Demand Sensing and Forecasting Challenges with External Demand Signals
According to the survey, the top demand sensing and forecasting challenges are related to issues with data: its collection, visibility, and analysis. It’s no surprise that all of these issues are directly tied to data inconsistencies. Clean data is essential to ensure accuracy and consistency, especially when integrating external datasets.
When asked to identify their top challenges in demand sensing and forecasting, respondents cited the following: no real-time visibility along the supply chain (41%), current processes are too manual (39%), inaccuracies in data within the organization, partners, and suppliers (37%), and no real-time sensing of demand and supply changes (34%). Understanding demand and supply shifts, and reacting accordingly, is at the heart of demand sensing and forecasting. From the demand side, shifts are the result of changing consumer preferences, brand loyalty, or economic factors. From the supply side, these market shifts are tied to raw material pricing or availability, labor shortages, or new entrants to the market. For those companies that cannot sense shifts in real-time, their forecasting accuracy suffers, thus leading to lost sales and higher cost of goods sold.
Supply chain visibility has been a hot topic over the last few years, but most people think of it only from a shipment standpoint. Point-to-point tracking solutions have seen billions of dollars in venture capital investments, but supply chain visibility goes well beyond these solutions. Supply chain visibility enables companies to track the location and status of products, components, and materials as they move through the supply chain. However, it also encompasses the entire end-to-end supply chain, from the sourcing of raw materials to the final delivery to the end consumer. At the core of supply chain visibility is access to real-time data for inventory optimization, tracking, and potential disruptions. To respond effectively to demand changes, companies must be able to adjust inventory levels quickly in response to market volatility and shifting consumer demand.
The second challenge identified by respondents is reliance on manual processes. More and more often, we hear about the autonomous supply chain. Automated demand sensing processes leverage real-time data and advanced analytics to predict short-term demand fluctuations, while manual methods rely on human interpretation of data, which can be time-consuming and prone to errors.
A third challenge highlighted by respondents is inaccuracies in data from within the organization, partners, and suppliers. As far back as 1957, computer scientists have referred to this as “garbage in, garbage out.” In a syndicated newspaper article about US Army mathematicians and their work with early computers, Army Specialist, William D. Mellin explained that computers cannot think for themselves, and that “sloppily programmed” inputs inevitably lead to incorrect outputs. A lot has changed since then, but the underlying principle is the same. Inaccurate data will lead to errors in demand sensing and forecasting, which will impact inventory management, supply chain operations, and profitability.
Demand Sensing and Forecasting Capabilities to Improve Forecast Accuracy
According to the survey, the capabilities respondents believe would most improve their ability to accurately forecast demand correlate with their biggest challenges. The top capability survey respondents said would improve their ability to forecast demand is the ability to ingest and analyze real-time data from many sources in disparate formats (27%). InterSystems Supply Chain Orchestrator is a data platform that ingests all relevant data from the sources that matter, both internally and externally, including geopolitical events, information on supply chain product integrity issues, supplier fulfillment discrepancies, and much more. Harmonizing and normalizing all this information to provide accurate data in real time, the platform simulates your business processes and then applies embedded AI and ML capabilities. With no “rip-and-replace” needed, companies gain accelerated implementation of powerful new capabilities, while lowering total cost of ownership in a way unmatched in the industry today.
The second capability identified by respondents is integrated inventory management with enterprise resource planning (ERP) and electronic point of sale (EPOS) to automate demand-sensing and forecasting (24%). Supply Chain Orchestrator enables organizations to adjust forecast plans with high levels of accuracy to successfully navigate sudden events, disruptions, or trends that affect demand, transforming fulfillment optimization. By leveraging demand sensing, organizations can increase output by adjusting production schedules in response to predicted demand, ensuring they meet customer needs effectively. Organizations can integrate more advanced sensing and forecasting capabilities with their point-of-sale, ERP systems, or applications, achieving faster time-to-value.
Final Thought on Demand Sensing and Forecasting
To be agile and competitive, organizations must be capable of extracting critical insights in near real-time. This remains a significant challenge when so many businesses lack end-to-end visibility or rely on manual data analysis and ad hoc provisioning and integration of different solutions. For demand sensing and forecasting, a reliance on manual data analysis, especially given the current state of disparate data streams, can be catastrophic. If companies are unable to understand the reasons behind supply shifts, they will be unable to adjust their demand forecasting accurately, which will lead to improper inventory availability, lost sales, and higher cost of goods sold.
Demand sensing and forecasting efficiency requires unified, trusted, and harmonized data. As an intelligent supply chain decision intelligence platform, InterSystems Supply Chain Orchestrator provides a complete view of an organization’s supply chain, harmonizing and normalizing disparate data from applications, suppliers, manufacturers, distributors, retailers, and consumers. It uses AI and ML to uncover what is currently happening, predicts what is likely to happen next, and uses prescriptive insights to outline the best options, ensuring maximum effectiveness and minimum delay.
Read the full report here.
Chris Cunnane is the Supply Chain Product Marketing Manager at InterSystems. In this role, he is responsible for developing and executing marketing strategy and content for the InterSystems supply chain technology suite. Chris has 20+ years of supply chain expertise, leading the supply chain practice at ARC Advisory Group, as well as holding various sales, marketing, and operations roles in the wholesale, retail, and automotive parts markets. He holds a BA in Communications from Stonehill College and an MA in Global Marketing Communications from Emerson College.
The post From Data to Decisions: Revolutionizing Supply Chain Management with Demand Sensing and Forecasting appeared first on Logistics Viewpoints.
