Demand Forecasting and Planning During Covid-19 Pandemic

Demand Forecasting and Planning During Covid-19 Pandemic

As the year 2020 draws to a close, the second wave of Covid-19 continues to rock the global marketplaces with a partial lockdown in major parts of the world. Almost all industries have been hit hard, as customer demand and inventory supply become more complex than ever. How can supply chain leaders and demand planners improve forecasting accuracy and ensure business resilience against future shocks?  

1. Flag Outliers for Improved Demand Forecasting

Coronavirus pandemic has triggered an abrupt shift in consumer behaviour, demand trends and sales channels. While essential services are in full swing, other sectors such as travel, hospitality and tourism remain in a deep slump. Even though some industries have begun to pick up, it is critical to monitor and minimize the forecasting models errors. One of the simple solutions is to flag these non-standard behaviours or “points outside the curve” as an outlier. Demand patterns under Covid-19 are likely to distort your inventory forecasts, but they should be treated as a one-off event that does not represent the level of the future demand. For example, the data of a massive spike in toilet paper demand in Q1 2020 will be irrelevant to forecast the seasonal demand in 2021. Hence, it is crucial to clean the baseline data and improve your machine learning models with forecast parameters sensitivity. By recognizing and monitoring these outliers appropriately, you can prevent over- or under-stocking and better manage your supply chain. 

On the other hand, flagging the Covid-19 lockdown months as outliers are not essentially the best option if some temporary demand shifts have become permanent. For instance, videoconferencing and online shopping under Covid-19 outbreak has seen a sharp growth and tend to persist into the future. It is therefore vital that the demand planners deeply understand their industry and timely detect the long-term changing customer behaviour. 

2. Identify the Drivers of Future Demand 

The forecasting models should take into account factors that influence the demand and what has changed due to the Covid-19 pandemic. Demand planners should be noted of emerging variables such as new customers’ preferences, advertising, promotion, market shares, price surges, supply chain disruptions, government restrictions, GDP.  Under the exceptional circumstances such as COVID-19, a best practice is to implement innovative Artificial Intelligence (AI) or Machine Learning (ML) analytical predictive models* that can understand the impact of those factors and provide accurate forecasts for the future. Companies should actively embrace advanced technology to navigate through the crisis and emerge stronger.

3. Form an Agile Planning Team

Companies should establish an agile team that brings together cross-functional expertise from data scientists, demand planners, sales, etc. to understand the rapidly changing situation. Frequent meetings will help gather timely, relevant data and sophisticated analytics on a global scale. Those insights such as virus spread, and other macroeconomic information are valuable to update scenario modelling and related forecasts. Inevitably, traditional forecasting tools like Excel cannot rapidly capture new demand and supply disruption data on a weekly or daily basis. Thus, companies should embrace advanced analytic models such as Forecasty.AI’s solutions that consider multiple sources of insights such as point-of-sale data, social listening, market research, and latest trends. In this way, business leaders can optimize opportunities for recovery over the near- and long-term horizons.

4. Apply Scenario Planning (What-If analysis)

In times of high uncertainty, companies need to implement scenario planning to have a clearer picture of the key demand drivers and the potential impact of future events. Dynamic driver-based forecasting models incorporated with different scenarios can deliver what-if answers in cycle times. AI/ML and big data can overcome human limitations and help executive leaders eliminate bias from their decision-making process. 

5. How to Forecast Demand with Forecasty.AI

Forecasty.AI helps global clients make smarter & more profitable business decisions through smart forecasting modelling, economics and analytics. We create highly accurate predictions by applying advanced AI/ML on established and new data sources via a simple drag and drop solution. Our solution automates forecasting tasks, ensures explainability and further allow scenario simulations to improve management decision making.

To help our customers plan their businesses and prepare future crisis management better, we provide scenario planning solutions specifically to tackle challenges from COVID-19.

Contact Us for Demo