Data quality playbooks
Investigate inconsistent, missing, duplicated, or stale information. Choose a scenario, then the decision you need to make.
Missing Fields
Required values are absent in some records, and the team needs to locate the first loss boundary before deciding how to repair or prevent the gap.
Inconsistent Identifiers
Records that should refer to the same entity use different identifier values or formats, making joins, deduplication, and downstream counts unreliable.
Timezone Mismatch
Events or date boundaries are interpreted in different time zones, shifting daily windows, ordering, or reported totals between systems.
Duplicate Records
The same logical record appears more than once, inflating counts or creating conflicting states when identity and delivery behavior are unclear.
Stale Sync
A downstream view or dataset stops reflecting newer source changes, leaving users with data older than the stated freshness expectation.
Mismatched Counts
Two views or processing stages report different totals for what appears to be the same population, and the counting boundary is unclear.
Pagination Gaps
Records disappear between pages of a listing or batch traversal because page boundaries, cursors, limits, or ordering do not cover the full result set.
Wrong Sort Order
A list or export is ordered differently from its documented rule, often around nulls, ties, mixed types, or a secondary key.
Import Mapping
Incoming columns or attributes land in the wrong destination fields, lose values, or apply defaults because the import mapping is ambiguous or changed.
Export Discrepancies
An exported file differs from the source view in records, fields, totals, formatting, or timing, leaving the intended export contract uncertain.