You have landed on our US/Global Site.

Finding A New Normal In Demand Forecasting

Due to the crisis, most generic demand forecasting models in place today are no longer as accurate as they used to be.

Covid-19 has led to drastic changes in how we do things, especially when it comes to making critical demand forecasts that can strengthen a business for the future or at least for the next 2 to 12 months.

Due to the crisis, most generic demand forecasting models in place today are no longer as accurate as they used to be and a relative approach has to be taken in order to find the “new normal” when compared to traditional historical patterns.


Demand forecasts are essential in performing many functions for businesses in creating immense value. However, the most important current operational plans no longer fit perfectly to the present situation as historical data presenting the past, bear little predictive power to reveal the future accurately.

From revenue forecasting which helps with expense planning, to the sales and operations planning forecasts that are the ideal for inventory and manufacturing plans, to even the most critical demand forecasts that determine workforce planning. The workforce is the lifeline of any enterprise and making sure there are enough people available at a particular time and location to do the required tasks in hand is very important. Not to mention they are not bored out of their minds or overworked to perform well.

We at Quinyx are continuously learning the range of possible outcomes in the current situation and the total possibilities of the Covid-19 impact, both in the short and medium term. And it is certain that companies will really need to update their forecasts and take a fundamentally new approach to creating them.

Today machine learning models form a critical part of advanced forecasting processes. It is important to know that each model is unique and so is the context for each specific business. These models usually consider historic business-as-usual patterns of demand. However, by training them with new data, new scenarios are possible. In assessing how the future demand might be, human judgment combined with automation is becoming increasingly necessary to create really accurate models.


Develop short-term scenarios
The first few months of this pandemic have showcased us some uncommon scenes such as temporary and compulsory shutdowns, quarantines, and social distancing, among many others. All these will have a great impact on the supply chain affecting labor demand, sales channels, and customer behaviours.

What you can do immediately:

  • Create scenario forecasts instead of point estimate forecasts

  • Decrease forecasting horizon to enable more flexibility

  • Test labour scheduling strategies to respond to the range of projections

The thing is companies are built around predictability, not change. They have found the right way to communicate with their customers by fine-tuning their response to deliver high levels of customer service. But a more temporal shift to change takes time. The speed of operational decision-making such as deciding how many people need to be staffed tomorrow, how many products will be distributed, how many shops will be allowed to invent a new service is still slow. Making these decisions manually is a struggle when you want to react quickly and effectively. When making new decisions, it is difficult to trust new data and new patterns that you see. You don’t know what you see is accurate and if the demand will stay flat for a while or stay zero.

That’s why companies will need to shift from manual labor planning to transparent labor modeling with the help of fast, data-driven accurate insights, especially during uncertain times. Due to events moving so fast, there might not be enough time to collect and review all the relevant data affecting your models, so it makes sense to focus on the few critically important ones during these anomalous moments. A good combination of mathematics and human judgment in decision-making form the order of the day.

Considering the uncertainty about Covid-19 and how different companies in the economy will respond to it, having scenario forecasts is essential. If one relies solely on point estimate forecasts, it could lead to strategies and actions that are not effective. It’s always a better idea to define 2-3 scenarios that build in a number of factors affecting demand (demand drivers), also assuming the impact on customers, supply chain, and regulation.

How to Begin?

Companies operating in different geographies can pay close attention to their own demand trends days or weeks ahead on the Covid-19 curve and possibly see how demand has shifted year over year. Some market research also helps to get a sense of how customer and channel mix may vary from one spot to another, and use automated judgment to adjust this in their supply chain.

Frame 9

For example: The demand in the packaging industry followed a very
similar slope in most countries and segments, offset by a few days, '
according to a McKinsey customer pulse survey report.

Decide what constrains your supply chain. For manufacturers – raw materials, production capacity, and your workforce will determine how well you can meet demand. Demand for flexible packaging, for example, is increasing greatly as it’s witnessing a supply chain disruption on account of voluntary shutdowns of factories or authorities enforced lockdown. Almost 70% of the packaging demand in Europe comes from the food and beverage segment. The luxury personal goods sector, home care applications, and other non-food sectors are expected to witness a decline in the Q3 and Q4 2020. Since these are combating the crisis, the supply chain cannot match it quickly enough, unless perfectly planned for.

In a competitive world, it’s more critical than ever to be empathic for customers and listening to their demands. In the process, you may discover the hidden treasures. Understand how well your distribution channels reach your customers and whether channel issues could cause them to stop buying from you. If so, have the right workforce and execution strategy to withstand any such challenges.

For example, there is still demand for apparel, but it is increasingly satisfied via e-commerce. Make sure you are not understaffed in your online fulfillment and distribution centers. Companies without multiple strong channels will need to curtail their demand expectations accordingly. These demands must be met by having enough staffing supply, something which can be determined using algorithms. What is important here to do is having forecasts which correctly reflect the business needs.

If your labour demand forecasts and corresponding employee shifts are created 4 weeks ahead, might be good to rethink for better business performance. Forecasting accuracy increases when the time horizon shortens and it becomes more difficult to make a forecast 4 weeks ahead, compared to 1 week ahead. Even something like the weather is influencing demand for some companies heavily, and due to its unpredictability, only short term forecasts can incorporate it.


Frame 7

Since forecasts become more accurate closer to the day of
operation, it makes sense to actually update forecasts with the
latest data available.

Once you have your forecasts, you can begin testing labor scheduling strategies to respond to the range of projections, making sure they are robust enough to deal with a range of possible outcomes. Know the objectives of each strategy. Do you intend to increase revenue, market share, profit, or willing to go off some of that by ensuring more customer fulfillment rate, customer satisfaction, or reduce costs primarily? In any case, ensuring employees are staffed well enough to meet these objectives is essential, and these can be taken into account when creating your staffing strategies.

Enhance medium-term scenarios
If we move from short to the medium term—the one or two quarters following the start of the recovery when economies have fully opened up and both the supply chain and sales channels return to their normal capacity, it makes sense to start with your pre-COVID forecast and then adjust scenarios accordingly.

Some things to keep an eye on:

  • Recovery Speed. Countries that are recovering faster can provide some sense of the time it will take to bounce back. Data from AutoNavi shows that traffic in major shopping districts in China, for instance, is picking back up, reaching 50% of its pre-crisis level in the first month of recovery. Similar trends are expected to follow Europe and North America.

  • The Hangover Effect. Some sectors such as supermarkets will experience a short ‘hangover’ especially if demand increased sharply during the crisis as customers stocked their houses (daily essentials like groceries, home care products). Other sectors like restaurants, bars, hotel chains may see a surge from pent-up demand.

  • Customer Behavioural Pattern Changes. Some temporary demand shifts may become long-term trends. Travel might decrease significantly if video-conferencing tools continue to shine. Or demand may stay normal but the channel shifts. So there could be a rise in the demand number of contact center agents if your online segment is increasing more rapidly compared to your physical segment, for example.


Once we are through with this the pandemic, we will all find ourselves in a new normal. Many companies will need to revamp how they forecast going forward, either due to a permanent shift in customer behavior or staffing model(s).

When rebuilding and putting the pieces back together, you will come across many opportunities to improve your forecasting approach and learn from that. It will be an ongoing process, as it always has been. Even if you can improve your labor demand forecasts from 63% to 87% it can have a great impact on your business.

Correcting demand forecasts completely affected by Covid-19 is quite emergent, but also completely doable if you have the right partners in place. With Quinyx’s dynamic monitoring of forecasts and labor demand algorithms, companies can quickly spot inaccuracies. Businesses can reasonably question how their demand drivers affect different aspects of the business. They can assess if the demand signals they are receiving from their immediate customers, both short and medium-term, are realistic and reflect underlying uncertainties in the forecast. We provide a granular view and help you reflect the accurate demand into accurate shifts that are automated and 100% compliant.


Accurate Demand forecasting is one of the main foundations of all data-driven decisions in the business process. It optimizes all sorts of aspects of the supply chain, especially distribution planning, production scheduling, inventory management, and strategic staffing. It lowers costs, improves services, and stimulates growth. With Quinyx, you will have it covered to measure your pathway to creating successful demand forecasts.


Book a demo 

Get in touch for a free walkthrough

Enter your details below and we’ll be in touch to book in a free, no-obligation demo with you. Is this a support related question? Please visit our Help Center.