Spreadsheet debt: the hidden operating system of manufacturing

A spreadsheet is the only tool in the plant that a business person can change on the day the business changes. ERP change requests take a quarter. The spreadsheet takes an afternoon. Every serious spreadsheet model in a factory is evidence of a real requirement that the systems of record did not meet.

This is why 'ban the spreadsheets' initiatives fail. The spreadsheet is not the disease. It is the immune response.

Key-person risk: one planner or one cost engineer holds the logic. Version drift: three copies of the same model with different assumptions circulate by email. Silent errors: a hard-coded cell from 2019, a broken lookup, a manual paste that skipped a row — and no test that would catch any of it.

Downstream, these errors propagate into scheduling engines and Dashboards. We have seen human entry errors in a master-data spreadsheet send a scheduling algorithm into an infinite loop; the visible symptom was 'the scheduler is slow', and the actual cause was three malformed rows.

The valuable part of a spreadsheet model is the logic: the business rules, the exceptions, the tribal knowledge encoded in nested formulas. The fragile part is the container — a file, on someone's drive, with no lineage, no validation and no audit trail.

The migration path that works is to lift the logic into a governed Flow where each step is named, testable and reusable, then leave the spreadsheet as an input or an output if people still want it. Nobody has to stop using Excel on day one.

List every spreadsheet a decision depends on. For each: who owns it, how often it is rebuilt, what feeds it, what it feeds, and what breaks if it is wrong for a week. Rank by that last column.

The top three are your data-product backlog. They are also, almost always, the fastest business cases to prove — because the manual rebuild time is already on someone's timesheet.