Modern industry is in no short supply of two things: metrics and moving parts. As commerce continues to evolve and grow in complexity, the metrics used to analyze and interpret that complexity seem to multiply by the day.
Of course, no two metrics are built alike; just like all statistically based measurements, they’re only insightful if used properly. To further muddy the quantitative waters, companies must also understand that metrics and models do not exist in a vacuum and are always evolving.
In other words, metrics work in conjunction with one another to build a complete story for a company to interpret. Anything short of an intersecting perspective on the quantitative side of a business is simply providing a snapshot, not a complete narrative. Fortunately, organizations can use Forecast Value Added (FVA) techniques to get the most from their metrics and analysis, trimming the proverbial fat from forecasting models to better inform their decision making.
Forecast Value Added: A primer
As defined by IBF, Forecast Value Added refers to “the change in a forecasting performance metric that can be attributed to a particular step or participant in the forecasting process.”
Forecasting models are among the most complicated of the quantitative endeavors a company can attempt. Given the seemingly infinite number of variables involved in forecasting, including consumer preference, economic climate, and even unknown factors like the weather, forecasting models are intricate, often cumbersome, and needlessly complex if too many variables are involved.
To that point, with so many moving parts, finance departments are constantly retooling and reconfiguring their forecasting models to further refine their abilities and results. One of the primary ways they accomplish such a daunting task is by measuring the FVA for the many factors involved.
In short, FVA is the amount a forecasting metric changes when a specific variable used in the calculation or process is altered or even removed. FVA is instrumental in that model-refining process, letting finance departments quantify the impact of each of the many factors involved and retool their models to become more precise and insightful. FVA defines the value and, therefore, the usefulness that each variable adds to the modeling process.
FVA in practice
To use a simple example of FVA at work, let’s take a hypothetical forecasting model that a company might use that has a mean absolute percentage error (MAPE) of 36 percent. After a meeting with the different department managers within the company, the finance department has been given a more refined set of variables to use in the forecasting model that has removed some unnecessary variables while improving the accuracy of others.
After making the recommended changes to the model, the new results yield a MAPE of 33 percent. In this example, the FVA of all the changes is 3 percent. If management wanted to reveal the FVA for each of the individual changes, the model could be run multiple times, incorporating a single change with each iteration to find the FVA for each of the removed or revised variables.
Obviously, given the importance of forecasting to the ongoing viability of any company, the need for precise, insightful data is critical to yielding reliable results that management can use in all aspects of the decision-making process. As a market leader in business data analysis, Vanguard Software is an invaluable partner in developing forecasting models needed to navigate the complicated waters of a crowded, ever-evolving marketplace.
About Vanguard Software
Vanguard Software introduced its first product for decision support analysis in 1995. Today, companies across every major industry and more than 60 countries rely on Vanguard Software’s Integrated Business Planning (IBP), forecasting and advanced analytic cloud platform. Vanguard Software is based in Cary, North Carolina.