Quick triage and prioritise (30–60 minutes)
- Create three short HubSpot lists to surface the noisy records: (a) Contacts/companies created or updated in the last 30 days, (b) Records with the target wrong value (e.g. incorrect country, blank company name), (c) Records where the same property changed in the last 7 days.
- Export a 200–500 row sample from each list to a spreadsheet. Look for patterns: identical timestamps, repeated placeholder text, identical user in the 'Last modified by' column, or identical import-like values (eg. "imported_contact_2026").
- Use simple counts in the sheet: how many records share the wrong value, how many were updated in a single hour, how many were created by the same source — those patterns point to bulk actions (imports, integrations) rather than one-off manual edits.
How to trace the likely source (structure your checks)
Start at the record timeline. Open a handful of suspect contacts and look for: "Form submission" events, "Property changed" events (who made the change), and "Imported" notes. The timeline usually tells you whether something was set by a user, a form, an import or an integration user account.
Next, check these HubSpot places in order: Import history (Settings > Data Management > Imports) to see recent batches and mapping; Forms > Submissions to find mis-mapped fields; Workflows to see bulk updates; and the activity feed on each record for API/integration user names. For Zapier/Make, use their run history to match timestamps and payloads against the record changes you exported.
Look for obvious signatures: many records changed at the same minute (import/bulk automation), identical placeholder text in a field (bad mapping), or changes by many different users one-by-one (manual edits). When in doubt, pick five records with the same error and trace each timeline — if they all show the same event type, that’s your source.
Prioritise fixes and six short prevention steps
- Blame list to decide what to fix first: (1) Recent large import batches, (2) High-volume integration runs (Zapier/Make scenarios), (3) Public form mappings, (4) Workflow bulk updates, (5) API scripts or middleware, (6) Manual edits across users. Triage by impact: number of records affected and business processes they break (reporting, billing, workflows).
- Short remediation steps to run now: export affected records, snapshot the wrong values, fix a small sample in a spreadsheet, re-import with "update existing" (mapped correctly) or correct the mapping and re-run the import; pause offending workflows/integrations while you fix mappings.
- Six prevention steps (apply in the next week):
- Add an import checklist (who, sample rows, mapping review) and require a reviewer for large imports.
- Lock or hide risky form fields, add clear labels and use dropdowns instead of free text where possible.
- Use a quarantine list: route new imports/integration updates into a temporary owner or property value and review before they hit live workflows.
- Restrict edit permissions for key properties and keep a short ownership list for who can run imports or edit mappings.
- Add a tiny audit automation: tag records updated by imports/integrations with a timestamp property so future triage is immediate.
- Standardise one 'source' or 'data origin' property across systems so every record shows where it first came from.
If you want a short, hands-on session to run this triage with your team, Optira can help with a focused audit and checklist.