Lineage answers two questions about any figure: where did this come from, and what depends on it?
Looking backwards, lineage traces a number on a dashboard through the transformations that produced it, the business rules applied along the way, and the source tables and files it was built from. When a figure looks wrong, this is how an analyst finds out why.
Looking forwards, lineage shows everything that depends on a given table or field. Before an engineer renames a column, replaces a data feed or changes a rule, lineage shows which reports and dashboards will be affected. This is often called impact analysis.
Lineage can be recorded at different levels of detail. Table-level lineage shows which tables feed which. Column-level lineage shows exactly which fields feed which, and is far more useful for debugging and impact analysis.
In pharma commercial analytics, where data arrives from many vendors and passes through many transformations before it reaches a scorecard, lineage is what turns “the numbers don’t match” from a week-long investigation into a short one. It also helps in audits, by showing how a reported figure was produced.