A polished chart can still be misleading when one connector stopped syncing. Data health should be checked before a performance change becomes a client conclusion.
Check freshness and coverage first
Freshness tells you when a source last synchronized. Coverage tells you how much of the selected period contains usable data. Both are necessary: a recent sync can still leave historical gaps, and complete history can still be stale today.
- Confirm the store and every active Ads source synchronized recently.
- Look for missing days inside the reporting period.
- Treat partial coverage as an explanation attached to the metric, not a hidden implementation detail.
Verify definitions before comparing totals
Check currency, timezone, revenue definition, refunds, and comparison dates. A midnight boundary or refund policy can explain a difference that otherwise looks like a tracking failure.
For Ads data, verify that the intended properties and campaigns are still assigned to the project and that the OAuth owner still has access.
Escalate with evidence
When a discrepancy remains, record the project, source, affected date range, last successful sync, and a correlation ID if one is shown. Do not send passwords, OAuth tokens, API keys, or raw customer data to support.