Glossary
Data provenance is information about where data came from and the activities and people involved in producing it. For a product attribute, provenance can identify the source document, relevant revision, extraction or editing activity and reviewer. It provides context for assessing evidence and responsibility, while leaving the accuracy of the source and interpretation open to verification.
A useful provenance record connects a value to a specific source and context. A general manufacturer homepage may not identify which document or revision supplied a particular technical fact. Record enough detail for another reviewer to locate the evidence and understand how it was used.
Where a value was transformed, inferred or manually corrected, distinguish that activity from direct transcription. The person responsible for review may be different from the person or system that entered the value.
A team extracts an operating-temperature range from a manufacturer datasheet. It records the document revision, the relevant table, the product model and the extraction date. A reviewer checks that the range describes normal operation rather than a storage condition or an absolute limit.
If the manufacturer later revises the specification, those references help the team identify which catalogue claims need attention. They also explain why a previous value was used.
Lineage focuses on the route through systems and transformations. Provenance gives broader context about origins and responsibility. A recorded source is evidence to examine, not a guarantee that a claim is correct or current. Human and AI-assisted enrichment should preserve that distinction when adding facts to product records.
Explore practical support for organising, improving and maintaining product information.