Reengineering an Ops Forecasting Model for a Global Mobility Intelligence Brand
Problem
A global mobility intelligence brand’s Operations Team had long utilized a manual, cumbersome, and often inaccurate forecasting model for shipments and installations, which informed revenue recognition and cost forecasting. The opaque system proved challenging to learn, explain, and utilize, and even more problematically was architected and operated by one individual, exposing a single point of failure.
When that individual departed, the group needed to build an enhanced forecasting model capable of scaling across multiple forecasters—thereby removing the single point of failure—and could easily be understood and explained to various stakeholders and executives. They brought in Excelerate to improve their processes and develop a clear, repeatable, and accurate model.
Approach
To develop a new forecasting model, Excelerate first had to understand the original one—a unique challenge considering the original model was littered with hard coded, unlabeled assumptions. This required reverse engineering, essentially tearing the model down to its individual parts to understand how it operated, where it went wrong, and how it could be improved upon.
Recognizing that accessibility and clarity were essential, Excelerate then developed an entirely new forecasting model around these needs. This required a deep understanding of how actual shipment processes work, how installations are scheduled, how revenue is recognized, and what caused delays in shipments and installations. Excelerate further pressure tested assumptions around curves, costs, backlogs, etc. to ensure inputs accounted for recent historical trends.
Results
Excelerate delivered the new forecasting model in the format the client requested—an enhanced, automated, and more transparent excel workbook—and additionally developed a python-coded model providing an easier-to-operate structure, especially when modifying assumptions and comparing scenarios. The excel version helped client stakeholders acclimate themselves to the new model and served as a foundation for the ultimate transition to python.
Excelerate validated the new model using input data from the previous (Q2 forecast), with outputs compared to the actual Q2 forecast PLUS actuals for shipments and installations from recent months. Excelerate further helped the client run their Q3 forecast using the new model to ensure adoption and accuracy. Finally, Excelerate provided visualizations to allow the Analyst Team to easily explain to the Executive Team why the new forecast may have changed from previous iterations. Given the downstream implications to revenue and cost forecasts, the Executive Team held high expectations around fully understanding variations in the model from previous iterations.
Excelerate helped the client mitigate its single point of failure, developing and deploying a clear and accurate forecasting model that could be easily deployed by any number of analysts in operations.