Product launch in 8 weeks. Three departments. Three different component counts.
Engineering BOM: 892 components.
Manufacturing BOM: 886 components.
Procurement BOM: 901 components.
Which one is right? All of them had errors.
The investigation took 2 weeks. Manually comparing 900+ line items across three Excel files. Finding differences. Tracing root causes.
Discovered:
- 12 components on engineering BOM not released to manufacturing
- 6 obsolete parts still on procurement BOM
- 18 quantity discrepancies across all three
- 4 duplicate part numbers with different descriptions
Built a BOM reconciliation system.
Three BOMs uploaded. System extracts: part numbers, descriptions, quantities, revisions. Compares across all sources. Flags discrepancies by category. Generates reconciliation report.
Now runs weekly. Catches discrepancies before they become launch delays.
Last quarter:
- 47 discrepancies caught early
- Zero BOM-related production delays
- Procurement accuracy: 99.7% (was 94%)
The three BOMs will never match perfectly. But now we know exactly where and why they differ.
How many versions of truth exist in your data?