Data Lineage

Data lineage is the provenance trail of a number: where it was born, every transformation it passed through, every system it visited, on its way to the dashboard a decision is being made on. When a regulator asks 'where does this figure come from?', lineage is the difference between an answer and an archaeology project.

What It Is

Data lineage records the journey of data through the estate: origin systems, every extract, transformation and load, the logic applied at each step, and the destinations — down to the specific report field. Lineage exists at two granularities: table-level (this warehouse table comes from these sources) and column- or field-level (this KPI is computed from these fields, transformed by this logic). Modern platforms derive lineage automatically from pipelines; without such tooling, lineage exists only as tribal knowledge, and departs with its keepers.

Why It Matters to You

Lineage has migrated from nice-to-have to load-bearing under three pressures. Regulatory: reporting obligations increasingly demand demonstrable traceability from reported figure to source — an auditor's lineage request is a deadline, not a curiosity. Impact analysis: before changing a source field, lineage tells you every downstream consumer that will break — the difference between a planned change and an incident. And AI governance: AI systems trained or grounded on data inherit its provenance; an AI output whose inputs cannot be traced is a governance orphan, whatever its accuracy. The executive test question: pick a number from a board report and ask for its lineage — the minutes or months it takes to answer is the organisation's data maturity, measured.

When You'll Meet It

In regulatory reporting, data-platform programmes, and AI governance. Ask for lineage on a current report figure — as a standing test, not a one-time audit.