Duplicate Detection & Merge
Bizily automatically identifies duplicate customer records and can merge them while preserving all history. Every merge is logged for complete transparency.
How Detection Works
Bizily uses multiple signals to identify potential duplicates with a confidence score.
Matching Criteria
Auto-Merge Threshold
Duplicates with 85% or higher confidence are automatically merged. Lower confidence matches are flagged for manual review.
Merge Process
When records are merged, Bizily preserves all data from both records.
What Gets Merged
- - Contact info: Most recent non-empty values are kept
- - Bookings: All booking history is combined
- - Conversations: All message threads are linked
- - Notes: All notes are preserved
- - Tags: All tags are combined
- - Custom attributes: Most recent values are kept
- - Stats: LTV, visit count, etc. are recalculated
Primary Record Selection
The record with more complete information becomes the primary. The secondary record is archived but its data lives on in the merged profile.
Data Protection
Merges are performed safely with full data preservation.
Transaction safety - Merges are atomic. If anything fails, nothing changes.
Full snapshot - Complete data from both records is saved before merging
Row locking - Records are locked during merge to prevent conflicts
Merge History
Every merge is logged for complete transparency and auditing.
Merge Log Contains
- - Date and time of merge
- - Which records were merged
- - Confidence score at time of merge
- - Reason for merge (auto or manual)
- - Full snapshot of merged data
- - User who approved (for manual merges)
Manual Review
Potential duplicates below 85% confidence appear in the review queue.
Review Queue
Go to Customers → Duplicates to see potential matches. Each pair shows the matching fields and confidence score.
Actions
- - Merge: Combine the records
- - Not a duplicate: Mark as separate customers
- - Skip: Review later
Preventing Duplicates
Real-time checking - When new customers are created, Bizily checks for existing matches
Import validation - CSV imports flag potential duplicates before creating records
Channel linking - Messages from different channels are linked to the same customer