Technology and Data
The Technology and Data domain tests how FP&A systems, data architecture, and controls produce reliable planning and reporting. Expect questions on the system stack, data quality, master data, automation, dashboard design, and the controls that keep numbers auditable.
The FP&A system stack
- ERP: transaction source (GL, AP/AR, payroll).
- Data warehouse/lake: stores integrated data; supports history and granularity.
- ETL/ELT: extracts, transforms, and loads data; ELT is increasingly common in cloud warehouses.
- EPM/CPM: planning, budgeting, consolidation, management reporting; holds versions and scenarios.
- BI/visualization: delivers dashboards and self-service analysis.
Data quality and controls
Data quality dimensions: completeness, accuracy, timeliness, consistency, validity, uniqueness. Master data-chart of accounts, cost centers, products, customers-must be governed centrally; mapping errors cascade into every report. Controls include access segregation, audit trails, version control, ERP-to-EPM reconciliations, and change management for report logic.
Automation and dashboards
RPA suits repetitive, rule-based reconciliations; APIs and scheduled jobs move data; AI/ML forecasting needs backtesting, explainability, and human review. Dashboards should use KPIs tied to strategy, correct chart types, a stated refresh cadence, and drill-through to source.
Worked example
A regional revenue dashboard ties to ERP by region, but totals are 3% below the EPM plan. You verify the ERP extract, then the mapping table: one region is mapped to the wrong parent in the hierarchy, and late postings after the close cutoff are excluded. The fix is at the mapping and cutoff-control layer, not a manual plug. This preserves the audit trail and prevents recurrence.
Traps
- Shadow spreadsheets that break the single source of truth.
- Comparing actuals and plan at different close dates or FX rates.
- Automating a broken process, which scales the error.
- Vanity metrics that look good but do not drive decisions.
- Assuming warehouse data is real-time when loads are batch.
- No audit trail or version control on report logic.