How widespread is ML-based revenue forecasting in FP&A?
I just graduated and landed a job as a junior FP&A analyst at a scale-up, where I'll have a lot of ownership from day one.
My question is : How widespread is the use of machine learning to model revenue drivers in a SaaS/FP&A context?
I had a work-study contract during my studies, so I've already applied these algorithms in a previous role (operational controlling). But since I'm moving into a new role, I'd like to know what the most common ways are to improve forecast accuracy and whether you've seen any more "creative" forecasting approaches.
One example I had in mind : a hybrid 3-statement model using machine learning outputs as short-term revenue inputs, then switching to a more classical, judgment-based forecast (bear/base/bull cases) after month 3.
Another specific example : model the churn rate using leading indicators (how the customer is using the product, frequency/volume of plan features) rather than a general assumption based on historical statistics.
Thanks for your help!