Demand Modeling and Forecasting, Explained

Wouldn’t it be nice if we could just read a special book and know the future before it happens? Unfortunately, we know from watching Back to The Future Part II that if such a book existed, it would lead to a dystopian future. Thankfully, Marty McFly was able to destroy Biff’s almanac and return peace to Hill Valley.

Similarly, demand forecasts and demand models are both forward-looking documents. But unlike Biff’s almanac, demand modeling predictions are usually fleshed out within complex spreadsheets!

So today we’re exploring the relationship between demand forecasting and demand modeling. We’ll look at the difference between qualitative and quantitative forecasting. We’ll review some of the primary forecasting methods. And we’ll identify one way to improve forecasts.

Why highlight demand modeling? Because as the supply chain recovers from post-pandemic challenges, forecasting the future will be essential.

Forecasts at the Forefront of the Post-Pandemic Recovery

Supply chain upheaval is becoming obvious across many market sectors. I can’t go a day without hearing about a product or manufacturing component that’s suddenly unavailable. When that happens, was it the demand forecast that fell short, or the demand model?

By now, we mostly understand what happened to disrupt the supply chain. Basically, as the pandemic began to ease and consumers resumed normal behavior, a tsunami of demand hit across markets. But that demand was met with a lack of shipping containers. And then lingering social distancing at West Coast ports restricted personnel. The driver shortage made trucks scarce, too. In reality, many challenging factors combined.

Then, just as the global economy seemed like it was really starting to rev its engine, the Ever Given got stuck in the Suez Canal! (You can read more about the challenges and risks facing the supply chain here and in this interview.)

How many forecasts around the world have been altered by unforeseen events? I can only imagine. By and large, supply chain leaders are adapting on the fly to this new environment. But I predict that supply chain resilience is one long-term trend that sticks around well beyond the post-pandemic supply chain crisis. Dual sourcing, near-shoring, better demand modeling, and added agility will help protect businesses from future disruptions.

After all, the future is always right around the corner. So let’s dive into the prediction tools that will help you match supply with future demand.

Demand Forecasting

First and foremost, forecasting is the process of predicting or estimating future events based on past data, present data, and analysis of trends.

Demand forecasting is an art and a science. It’s a science because there are certain methods and processes one follows in order to arrive at a forecast. It’s an art form because forecasting entails a level of prediction, guesswork, and intuition.

More specifically, this speaks to the two classic types of forecasting. One is qualitative. The other is quantitative.

Qualitative vs. Quantitative Demand Forecasting Methods

According to Investopedia.com, “Qualitative forecasting models are useful in developing forecasts with a limited scope.” These models rely on expert opinions and work best in the short term.

On the other hand, “Quantitative methods of forecasting exclude expert opinions and utilize statistical data based on quantitative information.”

The classic qualitative analysis is more opinion-based and includes these elements.

• Feedback from current customers: This encompasses any information directly from the customer about future demand.
• Surveys from potential end users: The surveys involve asking target customers about their intent over the upcoming period.
• Opinions of decision makers: This is the input from the executive team.
• Analysis of subject-matter experts: Trusted advisors provide their opinions.
• Delphi method: This method compiles and builds analysis from a panel of experts.
• Sales force estimates: This includes the projections of your sales team and/or individuals.

Meanwhile, quantitative analysis methods rely on data, time periods, and numbers. These include:

• Naive method: With this approach, you look at the previous period of time and assume the same number of things will happen again.
• Moving average: This looks at a number of periods and assumes the average number of those periods will occur.
• Exponential smoothing techniques: Similar to weighted moving averages, this method weighs numbers from certain periods differently than others. For example, this method can account for seasonality.
• Trend projection: This looks at the trajectory of a trend and extrapolates it into the future.

Demand Modeling

Demand modeling is different from forecasting in a few key ways. It’s worth noting that modeling involves forecasting. But modeling takes forecasting to a deeper level of utility. In essence, I think of demand modeling as the reverse-engineering of a demand forecast into an actionable series of if-then scenarios.

According to Toolsgroup.com, “Demand modeling works… from the bottom-up, as opposed to top-down. It breaks the demand components into a series of internal and external factors, the demand stream, and looks at how each impacts demand to predict future demand.”

What does that mean?

Say your team creates a forecast. Next, you ask the group, “To service the demand level we just forecast, what raw materials do we need to buy, what investments do we need to make, and what loans do we need from the bank?”

Demand Modeling: The Push-Pull Processes

Another way to think of demand modeling is as the initiator of your push-and-pull processes.

For example, demand modeling requires you to initiate analysis and budget for push processes. These include raw material planning and purchasing, capability planning, capacity utilization, resource planning, inventory planning, and inbound logistics planning.

But demand modeling also initiates analysis of pull processes. These include pricing, order management, packaging, distribution, and outbound logistics.

In reality, demand modeling is a great conversation starter. It’s a way to learn the ins and outs of your business and processes—and find blind spots. In short, demand modeling is a more sophisticated approach that facilitates and improves on classic demand forecasting.

Machine Learning and AI

Forecasting efforts are moving toward integrating machine learning and artificial intelligence tools. These tools have the ability to uncover trends and patterns that might not be obvious otherwise. All told, these tools can help limit the risk of overestimating or underestimating demand.

These newer technological tools should only get better. I predict they’ll certainly continue to affect the forecasting industry.

Mitigate Risk and Plan for Uncertainty

The usefulness of a forecast aligns with the accuracy of a forecast. Did anyone predict the Ever Given Suez Canal crisis? I highly doubt it. As Wikipedia puts it, “Risk and uncertainty are central to forecasting and prediction…. in some cases, the data used to predict the variable of interest is itself forecast.”

The data you use to forecast should be as current and accurate as possible. Step one toward streamlining and standardizing your data collection process is through digitization. Software leaders like Vector offer digitization tools specific to the transportation industry.

The basic premise is this: Going paperless streamlines day-to-day workflows for drivers and back-office employees. Meanwhile, digitization allows you to collect and store cleaner, better data for future analysis. In other words, if you want to get back to the future, start by checking out digitization.

What Else Are the Experts Saying?

Are you interested in reading more qualitative expert analysis of the supply chain industry? We’ve been doing a series of articles that include interviews with supply chain leaders about the state of the industry. Learn more about what the experts are saying.

This post was written by Brian Deines. Brian believes that every day is a referendum on a brand’s relevance, and he’s excited to bring that kind of thinking to the world of modern manufacturing and logistics. He deploys a full-stack of business development, sales, and marketing tools built through years of work in the logistics, packaging, and tier-1 part supply industries serving a customer base comprised of Fortune 1000 OEMs.