The supply difficulties during the COVID-19 crisis demonstrated that efficient material planning is essential for any company. It ensures that the right materials are available at the right time and in the right quantities, allowing production processes to run smoothly and avoiding supply shortages. A crucial component of modern material planning is forecast models, which help companies accurately predict future developments, such as material requirements. However, the value of forecasts relies on one key point: knowledge – knowledge of the past and the interdependencies within the system. Forecast models can generally be divided into two categories: qualitative and quantitative forecasts.

Qualitative and Quantitative Forecasts

Qualitative Forecasts

Qualitative forecasts are based on subjective assessments and expert opinions. They are often used in situations where there is insufficient historical data or when future developments are heavily influenced by non-quantifiable factors. The most common qualitative methods include:

  1. Delphi Method: Experts anonymously provide their assessments on specific questions. The collected opinions are aggregated and returned to the experts for revision until a consensus is reached. For example, experts might give their opinion on the future availability and price developments of raw materials.
  2. Market Research and Surveys: By directly surveying experts, suppliers, and other stakeholders, valuable information about future demand can be gathered. This method is especially useful for introducing new products or for rapidly changing markets. For example, suppliers could be surveyed about lead times and quality standards before launching a new product.
  3. Scenario Analysis: Scenario analyses develop various possible future scenarios based on different assumptions and evaluate their impact on material requirements. This method helps better understand uncertainties and risks and prepares for them. In procurement, different scenarios, such as changes in regulations or geopolitical events, could be analyzed to assess their effects on the availability and costs of necessary materials.

However, these methods are difficult to implement effectively in demand planning, especially for estimating the need for parts on a bill of materials. They are more suited for new products or market trends. Most companies, however, have access to their own, often untapped, historical consumption data and need only to use it wisely.

Quantitative Forecasts

Quantitative forecasts use statistical and mathematical models to predict future developments based on historical data. These methods are more objective and often provide more accurate and reproducible results. The most important quantitative methods include:

  1. Time Series Analysis: Time series models analyze historical data to identify patterns and trends and predict future values. Typically, consumption data from the past months and years are used to forecast the upcoming planning period.
  2. Causal Models: Causal models attempt to collect additional variables, identify relationships between all variables, and quantify them to forecast future demand. A classic example is the increased demand for ice cream on hot summer days.

Both models can be supported by advanced algorithms using machine learning. These algorithms use large datasets to detect complex patterns and relationships.

What Are the Benefits of Forecasts, and What Challenges Do They Bring?

It is in human nature to want to know the future. We bet on sports, insure our belongings, or predict the weather. So why not forecast material consumption? If we get our forecasts right, this brings many advantages, primarily reflected in lower costs:

  1. Improved Decision-Making: Forecast models provide data-based decision-making support, helping experts with strategic and tactical planning. They enable well-informed decisions regarding material procurement, production planning, and inventory management. Methods like business intelligence and advanced analytics offer deep insights into material flows and demand developments.
  2. Accuracy and Efficiency: Improved and supported decision-making enhances order accuracy and reduces the risk of overstocking and thus higher storage costs, as well as material shortages. In procurement, a forecast of annual stock requirements is often enough to determine purchasing volumes in supplier negotiations and secure quantity discounts.
  3. Optimized Supply Chain: Furthermore, improved forecasting positively affects the entire supply chain. Suppliers can better plan their production and logistics processes, leading to smoother collaboration and a more responsive supply chain. This reduces lead times and improves customer service quality.

In casinos, results from previous games are displayed next to roulette tables, giving the impression that if several red numbers have come up in a row, the chance of a black number is higher—the catch is that the probability remains the same. The challenge with forecasts lies in understanding the underlying data and relevant interrelationships:

  • Data Dependency: The accuracy of forecast models heavily depends on the quality and availability of historical data. Faulty, incomplete, or outdated data can lead to inaccurate predictions and undermine the model’s reliability. Therefore, continuous maintenance and updating of data are
  • Complexity: Implementing and using forecast models requires a certain level of understanding in fields like statistics, natural sciences, IT, and the specific area involved (e.g., products, industries, markets, geography).
  • Adaptability: Forecast models are based on historical data and specific assumptions. In the case of sudden market changes or unpredictable events, such as pandemics or natural disasters, the models may deliver inaccurate forecasts. This requires contingency measures (e.g., safety stock) or rapid adjustment and flexibility in the planning strategy.
  • Costs for Implementation and Maintenance: The introduction and ongoing maintenance of forecast models can be costly. This includes investments in suitable software solutions, as well as expenses for training and external consulting services.

The Surprise Bag That Isn’t One

Forecast models offer numerous benefits for material planning, particularly in terms of accuracy, efficiency, and cost savings. By using advanced algorithms and data analysis—the integration of artificial intelligence is a logical next step on this digital journey—companies can make informed decisions, better manage their inventories, and optimize the supply chain. However, the challenges and potential disadvantages must also be considered, particularly data quality dependency, model complexity, and associated costs. As such, the results of forecasts can be both positively and negatively influenced, depending on the effort invested. Unfortunately, as with most things, nothing comes from nothing. Ultimately, even human “gut decisions” are forecasts—but system-supported and rule-based methods eliminate potentially unfavorable emotions and, when used correctly, provide better and reproducible results.

In the upcoming parts of this series, we will take a closer look at specific approaches to material planning. The next article will focus on consumption-based forecasting, which relies heavily on actual consumption data to determine future demand. Stay tuned to learn how you can further optimize and adapt your material planning!