Key Forecasting and Supply Chain Terms to KnowWritten by Dag Liodden on Nov 4, 2019, 8:00:00 PM
As you start the journey toward finding the best forecasting software—and the most efficient, effective system overall—for your supply chain, it’s important to ensure you and your team have a solid grasp on the ins and outs of the available options. After all, as food supply chain professionals know all too well, “minor” details can have a major impact on the bottom line over time.
Just as it is imperative for your team to be on the same page about strategy and approach, it is vital that you “speak the language” to better evaluate the positive impact a new technology could have on your business. From the basic and commonplace (“artificial intelligence”) to the more complex (“consumption-based modeling,” anyone?), here’s a cheat sheet of some terminology you may come across as you research the best forecasting options for your food supply chain:
Glossary of Forecasting and Food Supply Chain Terms to Know
Advanced Analytics1: analysis of all kinds of data using sophisticated quantitative methods to produce insights that traditional approaches to business intelligence (BI) — such as query and reporting — are unlikely to discover. Advanced analytics include predictive, descriptive, prescriptive, and artificial intelligence techniques.
Artificial Intelligence (AI)1: a set of technologies that seek to mimic human ability to understand data, find patterns, make predictions, and recommend actions without explicit human instructions or programming changes. What distinguishes AI technology from traditional predictive and prescriptive analytics is its ability to learn patterns that traditional techniques cannot, or have great difficulty learning, such as processing or generating natural language. It can manage varied and complex data inputs and recognize deep, non-linear patterns that are often missed by traditional forecasting methods and the human eye.
Consumption-based modeling using multi-tiered causal analysis (MTCA): uses point of sale and/or syndicated scanner data to analyze consumer demand history, measure the impact of marketing programs, and model various scenarios to shape and predict future demand; the resulting consumption forecast is then applied as a leading indicator in forecasting shipments and sales orders.
Data Engineers1: these individuals make the appropriate data accessible and available for data scientists. Historically, this role resided in the IT group. But with more emphasis on leveraging analytics in decision making and managing a proliferation of data sources, many organizations are now dedicating full-time resources to this role to reside in the supply chain analytics group.
Data Scientists1: individuals who can extract various knowledge from data ranging from exploratory data analysis to building machine learning, statistical, and optimization models.
Data Visualization: the graphic representation of data. It involves producing images that communicate relationships among the represented data to viewers of the images through systematic mapping between graphic marks and data values in the creation of the visualization.
Demand Sensing: using detail consumer (downstream) sales data to recognize demand signals and measure/quantify the impact of demand shaping programs, such as marketing and business strategies, as well as external factors that influence consumer demand, including price, advertising, sales promotions, in-store merchandising, economic factors, seasonality, trends, product mix, new item introductions, competitive activity, weather patterns, social sentiment, etc.
Demand Shaping: influencing the demand of a product to achieve a desired goal. This can be done by increasing the price of a product, for instance, when its demand is higher than its supply, or by promoting substitute products.
Demand Shaping Actions/Influencers: internal programs and/or external factors which impact baseline demand, either positively or negatively. Examples include price promotions, clearance pricing, advertising, displays, new store openings, new distribution, weather, competitor activities, seasonality, shifts in holiday timing, etc.
Demand Signal Analytics (DSA)2: combines visual and predictive analytics to access the data in Demand Signal Repositories (DSR) to uncover insights. It starts with visual analytics to transform the DSR data into a format that allows for exploration, analysis and insight that suggests areas of focus, improvement and action. It identifies and measures market signals, then uses those signals to shape and forecast future demand automatically. Forecast outputs can be easily modified without programming, based on decisions you make relative to the demand shaping activities you select to incorporate.
Demand Signal Repository (DSR)2: a centralized data warehouse designed to store, organize, harmonize, integrate and cleanse large volumes of demand data and demand influencers for use by manufacturers/producers to efficiently serve retailers and consumers. Data includes: retailer point-of-sale (POS) scan data; syndicated retail scanner and consumer panel sources (i.e. IRI, Nielsen); wholesaler data (electronic data interchange, inventory movement, promotional data); customer loyalty card data; social sentiment data from social media mining; weather; holidays; seasonality; and other sources of structured and unstructured data that impacts/influences consumer demand. The primary focus has been on cleansing the data and synchronizing it with POS, syndicated scanner and internal (shipment and replenishment) data to enable companies to provide business users with a more complete view of retail performance. The repository itself is a database that stores the information in a format that allows for easy retrieval so that users can quickly query the database to identify what’s selling, where, when, to whom and in what quantities. It can become the foundation for a comprehensive information architecture strategy supporting an array of demand and supply-related predictive analytic applications and processes.
Descriptive Analytics: examination of data to report what happened (i.e. historical sales)
Diagnostic Analytics: examination of data to understand why something is happening (i.e. key drivers of sales performance)
Digitalization3: using digital technologies, tools, and information to change business models and operations, including automation, process efficiency and data transparency.
Digitization3: turning information into code so that computers can store, process, transmit, and use it for analytics and integration with other data sets.
Forecast Bias4: the tendency to deliberately over or under forecast demand to achieve a specific objective or disguise/accommodate for lack of confidence in forecast accuracy.
Forecast Error4: statistical measurement of the difference between the actual demand and forecast demand. Typically expressed as Mean Absolute Percent Error (MAPE) or Weighted Mean Absolute Percent Error (WMAPE).
Machine Learning5: machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and uses it to learn. The process of learning begins with observations or data, such as examples, direct experience, or instruction, in order to look for patterns in data and make better decisions in the future based on the examples that are provided. The primary aim is to allow the computers learn automatically without human intervention or assistance and adjust actions accordingly. Advanced machine learning algorithms are composed of many technologies (such as deep learning, neural networks and natural language processing), used in unsupervised and supervised learning, that operate guided by lessons from existing information.
Predictive Analytics: statistical techniques that leverage data to project what is likely to happen in the future. (i.e. sales forecasts over a future time horizon)
Prescriptive Analytics: statistical techniques that leverage data to recommend actions based on decisions, constraints, etc. (i.e. optimized production plan and schedule)
Sales & Operations Planning (S&OP)6/7: a structured, regular process that aligns all functional areas under a unified set of assumptions to enable and coordinate decision making. It integrates demand, supply, operations, and financial planning into one game plan for business. It also links strategic plans to operational plans and attempts to provide a decision-making framework to maximize sales and profit. A key concept of S&OP from a demand perspective is that the resulting plan is what the Sales & Marketing organizations commit to deliver. The process starts with an unconstrained demand-driven forecast and builds to a consensus demand plan that sets the overall level of supply chain operations output (manufacturing, purchasing, etc.) and other activities to best satisfy the sales forecast while meeting the general business objectives of profitability, productivity, customer satisfaction requirements, etc.
Structured Data8: clearly defined data types, typically text only, whose pattern makes them easily searchable and resides in relational databases and data warehouses. It is relatively easy to analyze using existing analytics tools. Examples include retail scan data, shipment history, inventory levels, etc.
Unstructured Data8: data that is usually not as easily searchable, including formats like audio, video, and social media postings. The data is structured internally to its source, but not structured via pre-defined data models or schema. It may be textual or non-textual and human or machine generated. Examples include email, websites, mobile texting data, media, MS Office documents, satellite imagery, sensor data, digital surveillance images, loyalty card data, social sentiment data (social media), consumer perception/attitudinal research, online and mobile shopping records, RFID, IoT, etc. There is much more unstructured data than structured; unstructured data makes up 80% and more of enterprise data and is growing at the rate of 55-65% per year. Without the tools to analyze this massive data, organizations are leaving vast amounts of valuable data on the business intelligence table.
1 Use Gartner’s Five-Stage Maturity Model to Reach Supply Chain Analytics Excellence, Gartner, August 2, 2017
3 Digitization, Digitalization, And Digital Transformation: Confuse Them At Your Peril, Jason Bloomberg, Forbes, April 29, 2018
4 Institute of Business Forecasting & Planning; Tackle Forecast Bias by Driving True Accountability into S&OP Decision Making, Gartner, February 5, 2018
Subscribe to our blog to stay current on the latest and greatest in the field of demand forecasting.