Comparing a human-generated forecast with a predictive forecast within Corporate Performance Management (CPM) software involves the capability to align manual projections created by business users with data-driven predictions generated by the system. With this capability, decision-makers can acknowledge any deviations between subjective estimates and objective, predictive insights, offering a more balanced, cohesive view that can enhance strategic planning and execution.
Scenario: A manufacturing enterprise uses CPM software for production planning. The sales team makes a human-generated forecast based on their knowledge and experience, while the CPM software system generates an AI-driven predictive forecast based on historical sales data, current market conditions, and various other factors.
Solution: The software's capability to compare these forecasts helps identify any major disparities and commonalities. This comparison provides valuable insights into potential opportunities, risks, and uncertainty levels, assisting in making more informed production decisions.
A human-generated forecast may include an extensive driver-based model which was configured by finance to take sales data and calculate a forecast. We still consider that “human-generated” for the sake of this requirement.
As always with predictive analytics, historical data is required to generate a forecast. Make sure you have at least 2 years of data to make the most out of this functionality.
Being able to put these side by side and spot major deviations is crucial. This might include a simple set of columns with a separate column showing the difference between, or a visualization with traffic lighting highlighting major deviations from the human vs predictive models. Traffic lighting works great, as it quickly distills the analysis and discussion down to just a few data points, allowing the sales team to explain why their model has an exception to the rule built into it.