Demand forecasting and uncertainty: What has changed in recent years?

Is the world today a more uncertain place than it was 10, 20 or 30 years ago? Probably so. Or at least that is the general feeling. The beginning of this era of ‘uncertainty’ was the COVID-19 pandemic, a global event that has had the greatest impact on people’s lives over the longest period of time in at least the last half-century.

This sea of uncertainty has hit supply chain planning hard and, inevitably, one of its cornerstones: demand forecasting. In this post, we examine what has changed when it comes to demand forecasting.

The evolution of demand forecasting in recent years

Demand forecasting has always been a key activity for the supply chain, but recent years have completely transformed the way it is approached. Whereas companies once relied mainly on historical data to draw up their forecasts, they now operate in an environment where disruptions are increasingly frequent and difficult to anticipate. The pandemic, geopolitical conflicts and tensions in the supply of raw materials have shown that demand behaviour can change drastically in a matter of days.

Faced with this reality, planning processes have had to become more agile. During the COVID-19 pandemic, for example, many organisations – particularly those in the hospitality sector – moved from producing monthly forecasts to reviewing them weekly or even daily in order to adapt to a constantly changing market.

This shift has also had an impact beyond forecasting itself. Uncertainty over supply and rising raw material costs have led many companies to increase their stock levels to ensure product availability. However, excessive stock levels are not a sustainable long-term solution due to the additional investment, risks and pressure on the supply chain that they entail.

The difference between structural and cyclical changes

As we have seen, we live in an era of constant change… But not all changes are the same.

One of the greatest challenges in demand forecasting is determining whether a market shift is temporary or permanent. Whilst cyclical changes are usually linked to specific events and tend to fade over time, structural changes bring about lasting alterations to consumer habits, sales channels or customer behaviour.

The pandemic provided numerous examples of both phenomena. Some variations in demand were directly linked to the restrictions in place at the time, but others ended up becoming established. The growth of e-commerce is probably the most obvious example. What initially appeared to be a response to an exceptional situation became a permanent change in the way people shop. Similar transformations also took place in other sectors. In the drinks market, for example, daytime consumption and the so-called ‘tardeo’ (afternoon drinking) came to the fore, altering patterns that had remained stable for years.

For planning teams, the key lies in identifying when these changes begin to be reflected consistently in the data. When purchase frequency, consumption habits or demand patterns change consistently over time, it is no longer prudent to expect a return to previous behaviours. In such cases, demand forecasting must adapt rapidly to the new reality, drawing on both data analysis and business knowledge to make decisions with greater confidence.

Historical demand is no longer a reliable basis for demand forecasting

For decades, historical data has been the primary reference for drawing up a demand forecast. However, in today’s rapidly changing environment, historical data can no longer be used in isolation. Peaks in demand caused by exceptional circumstances can distort future forecasts, particularly when consumer habits or market dynamics have changed permanently.

For this reason, demand forecasting must be complemented by information from sales, marketing and other teams in close contact with customers. Understanding what is happening in the market and how competitors are evolving is just as important as analysing the data. When a trend persists for several years, companies must recognise that they are facing a structural change and adapt their strategy, whether by adjusting their product portfolio or investing in new lines of business with greater growth potential.

S&OP for a consensus demand forecast

The accuracy of a demand forecast depends on more than statistical models alone. For the forecast to truly reflect the market situation, it is essential to incorporate insights from areas such as sales and marketing. This is precisely one of the main objectives of the S&OP process: to build a shared view of demand that enables the entire organisation to align around a single plan.

However, achieving this requires more than simply implementing a methodology. It is necessary to explain the purpose of the process, foster collaboration between departments and secure the backing of senior management. Having internal champions and the support of senior management is essential to drive the cultural change required for S&OP and to make demand forecasting a shared responsibility across the organisation.

The customer at the centre, regardless of sector

Although processes and challenges vary across industries, the principles underpinning good demand forecasting are universal. Whether in B2B or B2C environments, the objective remains the same: to understand customer needs and ensure that products are available at the right time. The pace may differ, but the importance of delivering a high level of service remains constant.

Experience across different sectors also shows that every supply chain has its own specific characteristics. Whilst in the electronics sector the priority may be to secure components whilst maintaining minimum stock levels to avoid production stoppages, in the fast-moving consumer goods sector, distribution and the last mile take on particular importance. In sectors such as lubricants, the challenge lies in striking a balance between availability and costs, whilst maintaining optimal stock levels to meet demand without creating excess inventory.

The evolving role of the demand planner

And demand forecasting would not be possible without the demand planner, a role that has gained prominence as demand forecasting has become a more strategic process for companies. What was once a responsibility shared between sales, marketing and data analysis is now a specialised role that is usually integrated within the supply chain or operations function. Their role is no longer simply to produce forecasts, but to bring together information from different functions to build a forecast that reflects the reality of the business.

For this reason, a good demand planner needs much more than just analytical skills. They must understand the market, collaborate with sales, finance and operations, and possess the curiosity needed to question the data and understand what lies behind it. The ability to interpret trends, anticipate change and assess the impact of decisions on the business is what distinguishes a report generator from a demand forecasting specialist who adds real value to the organisation.

The role of technology in demand forecasting

Technology has become an indispensable element of any demand forecasting process. As the volume of data and the complexity of supply chains increase, managing all this information through spreadsheets alone becomes increasingly impractical. Specialised tools enable data to be consolidated, calculations to be automated and key indicators to be visualised, facilitating faster and more informed decision-making.

However, technology alone does not guarantee a good forecast. Systems can process large amounts of information, but they still require business knowledge to interpret risks, opportunities or changes that are not yet reflected in the data. For this reason, the best decisions usually arise from a combination of advanced tools and the expertise of planning teams.

KPIs for measuring a good forecast

The accuracy of a demand forecast is usually measured using indicators such as forecast accuracy, but assessing the quality of a forecast requires a broader perspective. A forecast may appear correct on paper yet still lead to inventory issues, excess stock or lost sales. It is therefore important to also analyse metrics such as stock levels, obsolete products and the impact on working capital.

Ultimately, a good forecast is one that helps drive better business decisions. If the company maintains appropriate service levels, controls its stock levels and adapts its forecasts to market changes, this is a sign that the process is working. Furthermore, it is worth remembering that demand forecasting is not an exact science, but a skill that improves with experience, continuous learning and constant review of results.

The big question: Who is responsible for demand forecasting?

Although the demand planner plays a central role in the process, responsibility for demand forecasting should not rest solely with this individual. Their role is to facilitate the process, coordinate the different departments and transform the available information into an agreed forecast.

However, sales and marketing teams have the strongest understanding of the market and the customer, so their involvement is essential to building a reliable forecast.

For this reason, more mature organisations view demand forecasting as a shared responsibility. Linking forecast accuracy metrics to the sales teams’ targets can help reinforce this commitment and foster greater collaboration. Ultimately, striking a balance between service levels, inventory and working capital depends on all departments working in alignment with the same forecast.