Correct data is not the same as trustworthy data. A value can be complete, current and technically right - and still impossible to explain. In industrial environments, that gap has a name: traceability. The moment a figure is challenged - in an audit, a customer review, a root-cause analysis - “it looks right” is not an answer. What earns trust is showing where the number came from, how it was derived and what it refers to, without rebuilding it by hand.
IN BRIEF
- A value can be complete, current and technically correct and still be impossible to explain or defend.
- The missing property is traceability (data lineage): source, transformation, definition, and what the figure refers to.
- Calculation logic that lives only in one engineer’s spreadsheet leaves the company when they do.
- Traceability is built in by linking data to its source - not bolted on as a report at the end.
Part of the series, The Industrial Data Foundation
The situation
A correct number can still raise questions
Within any single system, the data is often perfectly valid: a machine value is current, a batch record complete, an emissions figure calculated by an internal rule. But the moment that value crosses a system boundary - exported, transformed, aggregated, dropped into a spreadsheet or a dashboard - even a technically correct number starts to raise practical questions.
One reported figure - the questions it has to answer
| Where did it come from? | which source system |
| Has it been transformed? | filtered, converted, aggregated |
| Which definition was used? | whose rule, which version |
| What does it refer to? | asset, batch, product |
| Can we defend it in an audit? | only if the path is visible |
A value you cannot trace is a value you cannot defend - however correct it happens to be.
Why quality isn’t enough
Complete, current - and still unexplained
Improving data quality is necessary, but it does not automatically create trust. Origins are unclear, transformations are hidden, different departments trust different sources for the same object, and the calculation logic lives in spreadsheets or scripts rather than a visible process. Too often the only complete account of how a figure is really produced lives in one long-serving engineer’s workbook - and it walks out of the door the day they retire.
The fullest account of how a number is built often lives in one person’s spreadsheet - and leaves when they do.
The way out
Keep data linked to where it came from
Traceability is not a report you add at the end; it is a property of how data is brought together in the first place. Keep the connection to the source, and the explanation comes for free:
- Link, don’t copy. Data is referenced live at its source, not copied forward until no one remembers which copy is authoritative.
- Referential integration. Values stay connected through meaning - assets, orders, batches, quality events - not flattened into isolated exports.
- Time as its own dimension. A figure can be reconstructed as it stood at the time, even after tags, units or hierarchies later changed.
- Audit trails. Origin, transformation and use are made visible, ready for a review or a customer question.
Traceability turns data from a technical asset into an accountable one.
So what now?
Start with one number that matters
Traceability does not have to begin as a governance programme. Pick one figure people already argue about - a yield, an emissions number, an OEE - and make it traceable end to end: link it to its source systems, capture how it is calculated, and record the definition. Once one number can be explained on demand, the same pattern extends to the next, and the reasoning that used to live in a spreadsheet becomes part of the system landscape.
The payoff is concrete. Audits and customer reviews stop triggering a scramble. Root-cause analysis starts from a visible data path instead of a manual reconstruction. And the knowledge no longer retires when a key person does.
The lesson travels
If you can’t explain the number, the gap isn’t quality
If your data looks complete and correct but is still hard to explain, validate or defend, the missing piece is probably not quality - it is traceability. Making industrial data explainable from source to use - the same foundation that keeps production numbers reconcilable - is exactly what the boTec DataBridge is built to be.