From "multiple truths" to audit-ready metrics

The Audit usually triggers panic. "Where did this numbers come from? Can you reproduce it?"

In an Excel-based world, the answer is often: "Ask Steve, but he's on holiday."

The Principles of Audit-Ready Data

To make your metrics defensible, you need three things:

1. Reproducibility: If I run the code today for "January 2024", I get the same number as I did last year.
2. Traceability: I can trace the number back to the raw invoice line item.
3. Isolation: Logic lives in code (SQL/Python), not in cell formulas.

The Playbook

Step 1: Get logic out of Excel

Excel is for display, not transformation. Move your SUMIFS and VLOOKUPS into SQL views or a transformation layer like dbt.

Step 2: Version Control

Store your metric definitions in Git. "Revenue v2.sql" with a commit message "Updated for IFRS 15 change" is an audit trail. "Revenue_FINAL_v3_REAL.xlsx" is not.

Step 3: Snapshotting

Data changes. Customer addresses change. For audit readiness, you must snapshot your dimension tables. "What was the sales region of this customer at the time of the booking?"

The Payoff

When the auditor asks for a sample walkthrough, you don't panic. You show them the lineage graph and the Git history. It turns a 2-week stressful deep dive into a 2-hour meeting.

---