Journal of Business Forecasting Articles Issues

During the current pandemic and subsequent demand disruption, we are preoccupied with understanding current consumer behavior and short-term demand. But equally as important is the longer-term picture. In this article I present a framework for understanding how demand for different products will shift, both as we exit this crisis and for the year ahead. I use a categorization system based on product needs, revealing how consumer behavior now and going forward will impact your demand forecasts.

This article discusses the importance of establishing strong links between demand planning and inventory management processes. Several steps in the Sales & Operations Planning (S&OP) process can be more effective if this relationship is managed well. For example, linking expected demand patterns with a review of stocking risks and an evaluation of the best stocking strategy. It shows “how-to” examples for integrating demand planning with inventory control to gain a better overall balance across objectives leading to improved financial performance.

Demand planning for the e-Commerce channel presents some unique challenges. Success in planning for this channel requires a different perspective on planning and a different approach to managing the planning process. It may also require a different organizational structure. In this article I share some of my experiences with planning demand for this channel, with the hope that others that are new to planning for this channel can avoid some of the mistakes I made when I first started.

This article describes a 5-step process about how the philosophy of Lean Manufacturing can be applied to demand planning. The steps include what customers truly value (valuable products being those customers are willing to pay for); mapping the value stream (i.e., how a product is delivered to the customer); evaluating opportunities for improvement; working on activities that are requested by customers; and aiming for continuous improvement.

Rapid demand response forecasting techniques are forecasting processes that can incorporate key information quickly enough to act upon in real time by agile supply chains. This makes it the ideal approach to plan demand during the current disruptions caused by Covid-19. Here I present a case study of using these techniques to assist a major online grocer, employing machine learning and advanced analytics to better predict demand. Combining product attribute data, a range of external data, and historical demand proved to be the best approach to predicting changing demand patterns, protecting this company at this difficult time, and giving it a powerful competitive advantage.

Using means and standard deviations of statistical forecasts has been the default method for demand planners for decades, but there is a key shortcoming with this approach, namely that it assumes that demand is normally distributed, which it rarely is. This incorrect assumption severely impacts forecast accuracy and accuracy of all dependent plans. The solution, and an increasingly adopted method, is probabilistic forecasting. In this article I discuss how probability distributions allow planners to work with the real uncertainty in demand and enjoy more accurate demand plans as a result. I also explore other benefits of this approach and the differences between deterministic and probabilistic forecasting.

It seems intuitively obvious that the companies who figure out how to best engage with their consumers will get more than their fair share of growth. As a result, integrating consumer demand into the demand forecasting and planning process to improve shipment (supply) forecasts has become a high priority for many companies. Most supply chain professionals are quickly realizing that their supply chain planning solutions have not driven down costs and have not reduced inventories or speed to market. Consumption-based modeling using a process called, “multi-tiered causal analysis” (MTCA) which links consumer demand to supply (downstream data to upstream data), using a process of nesting advanced analytical models. Although this process is not new in concept, it is new in practice. Consumption-based forecasting using the MTCA approach is a simple process that links a series of causal models through a common element (consumer demand) to model the push/pull (sell in/ sell out) effects of the supply chain. It is truly a decision support system that is designed to integrate statistical analysis with downstream (POS and/or syndicated scanner) and upstream (shipment) data to analyze the business from a holistic supply chain perspective.

Programmatic advertising allows companies to advertise to target markets and understand demand drivers and consumer behavior like never before. What’s more, this kind of advertising allows demand planners to shape demand in near real-time. By connecting programmatic advertising with demand and supply planning, we can use it to boost demand at will. It represents the next stage of demand planning and analytics, and offers forward-thinking companies the opportunity to gain a competitive advantage in the years to come.

There’s a lot of excitement lately about AI, new models, and machine learning algorithms and the accompanying idea that they will replace all human judgement. This misconception may be due to lack of understanding about how all the tools and methods now available fit together, and how we need all of them if we’re to forecast all datasets accurately. In this article we will look at the full spectrum of forecasting methods from pure judgment to machine learning, and classify each of them so that they are easy to understand. I also provide an explanation of each of the broader classes of methods, so demand planners can add different models to their toolkit, knowing when to use which one for maximum effect.

The purpose of demand planning is to improve forecast accuracy and increase supply chain effectiveness. Using current and past forecast performance, demand planners can provide guidance on what adjustments are needed to improve future forecast accuracy. The next step in developing an effective demand planning process is to make the future forecast more forecastable, that is, to take actions that make it more likely that the forecast will be accurate. While there will always be factors outside our control that will influence forecast reliability, there are actions we can take to increase the odds that our forecasts will be accurate. In taking these actions, we move from merely planning demand to managing demand.

Most forecasting approaches are based on the idea that a forecast can be improved if the underlying factors are identified then forecasted separately. It stands to reason, then, that with conventional time series approaches that extract information from historical observations, we may gain insights by isolating and modeling elements like seasonality. In this article, I explore how transforming historical data with multi-cluster time series analysis allows us to gather additional information about seasonality. I discuss a newly introduced modelling approach that combines information from many different clusters of attributes and levels of aggregation for improved seasonal profiles, and consequently allows us to build more robust forecasts.

In the late 1960s and early 1970s, Kahneman and Tversky carried out a series of breakthrough studies that gave rise to the field of behavioral economics, and changed how judgment should be viewed. Few of us are aware of the significance of Kahneman and Tversky’s findings, namely that human judgment is flawed and, in terms of making predictions about the future, mathematics tends to outperform human judgment close to 100 percent of the time. They used data and analytics to find true patterns in human behavior that replace false assumptions in a number of areas of study. Similarly, demand analysts and planners need to find the true patterns in human judgment to replace the false ones that govern our demand forecasts and plans.

There has been some discussion of late in IBF circles about Quick Response Forecasting (QRF). In my attempt to understand the concept, I have read a number of articles only to find myself a bit confused by proponents who discuss everything from Big Data, to point-of-sale data (POS), to “responsive” Supply Chains—each already having a place and a role in Supply Chain Planning, but none of which are conceptually new. On its surface, QRF seems to be a demand planning idea without particularly well-defined boundaries or purpose. After digesting the available literature, I think otherwise well-meaning folks are trying to define a process that already exists inside the Sales and Operations Planning process (S&OP). It is referred to as Demand Control (and sometimes SOE – Sales and Operations Execution) , and it is the transition point from more strategic planning to execution planning within S&OP.

The major challenge demand planners face is how to improve forecast accuracy in a way that is aligned with the company’s processes and strategy. Here are 10 key points to consider that will help in achieving this goal.

Quick Response Forecasting is a forecasting process that can incorporate information quickly enough to act upon by agile supply chains. The challenge most companies are experiencing is the length of time it takes for the demand forecasting process to incorporate rapid changes or short-term spikes in demand. Many supply chains are still too sluggish, and not agile enough to take full advantage of demand forecasts that include short-term spikes. Forecasting at the edge for real-time demand execution using event stream processing (ESP) combined with machine learning (ML) to analyze demand as it is generated at the store device can eliminate the noise decreasing latency in the decision-making process.

The article shows how Cisco improved its operational and financial efficiencies with PID (product ID) rationalization. Not only that, it outlines the strategy used to get the executive buy-in, and then implementing it. Savings can come from various sources. It shows all the sources of saving for a hardware (HW) portfolio, and the way to monetize them to know exactly how much has been saved from rationalization.