01 — Global systemically important bank
Rationalising a forecasting model estate
Situation
Years of local model-building had left a global FP&A function with overlapping forecasts, inconsistent metric definitions across business lines, and an administrative load that crowded out forward-looking analysis. Model risk was rising with the count.
Approach
- Inventoried the model estate and tested each model against the decision it was built to serve
- Standardised financial data intake, ETL and aggregation on the ERP/EPM reporting platform
- Established a common metric and definition layer usable across business teams
- Led a cross-functional team of 25+ across Finance, Risk, Products and Technology
- Redirected released capacity into scenario and driver analysis
70%+
Redundant models retired
20%+
YoY capacity shift to analysis
25+
Cross-functional team led