← Back to glossary
DataUpdated 2026-07-06

Deduplication

The process of identifying and merging duplicate records so that each real-world entity is represented only once in a dataset.

In more detail

Deduplication is the practical cleanup that follows from entity resolution: once you know several records refer to the same entity, deduplication merges them into one. Duplicates creep in constantly, from imports, manual entry, and multiple systems, and they distort counts, split history, and cause the same customer to be contacted twice.

The hard decision is what to keep. When duplicate records disagree on a field, which value wins? Merging has to reconcile the conflicts, preserve the history that matters, and avoid destroying good data. This is why safe deduplication is more than deleting near-identical rows; it is a careful merge with rules about precedence.

Where this shows up at Ceven

Ceven can run deduplication as a workflow, identifying duplicates and merging them under clear precedence rules, with human-approval gates on the ambiguous merges rather than destroying data automatically. It acts on the customer's own systems rather than being the record store, and every merge is captured in the audit trail so it can be reviewed or reversed.

Related terms

See it in production.

Start free