All docs pages (49)
Duplicate Finder
Find groups of likely-duplicate records by comparing text, email, phone or URL fields, then merge each group or keep one record and delete the rest.
Updated 2026-11-09
Duplicate Finder looks for records that were probably entered twice, such as "Acme Inc." and "ACME, Inc". Finding duplicates only reads data. Resolving a group is a separate step that changes your org: it merges the records or deletes the extra ones. Open Data → Duplicate Finder.
How to use it
- Pick an object in Object.
- Under Match on field(s), tick the fields to compare. Only text, text area, email, phone and URL fields are listed. Name is ticked by default when the object has one.
- For each ticked field, choose Fuzzy (a 0 to 100% similarity score, Jaro-Winkler, so small typos still match), Exact (identical after cleanup) or Phonetic (same Soundex code, so they sound alike). Cleanup lowercases values and removes accents and punctuation.
- In Match when, choose Average similarity ≥ threshold, All fields match or Any field matches.
- Set the Fuzzy similarity slider, from 60% to 100% (default 85%).
- Tick Company names (ignore Inc/LLC/Corp…) to drop endings such as Inc, LLC, Ltd, Corp and GmbH before comparing. It is on by default for Account.
- Click Find duplicates.
A field blank on both records is skipped for that pair. Groups chain: if A matches B and B matches C, all three land in one group.
What you see
The header shows Groups, Affected records, Scanned and Object. Each group shows its record count and a % match, which averages the similarity scores of matching record pairs in the group. Highest matches come first. Each group lists the record Id and your match fields, with a link to open each record in Salesforce. With no matches, the page says No duplicates found.
Resolving a group
Click Resolve on a group. The window re-reads the records and shows:
- Choose the survivor: the record to keep, marked KEEP. By default, the record with the most filled-in fields.
- Conflicting fields: each field whose value differs. Pick the value the survivor keeps.
- A preview, for example "Keep 1 record, merge 2 · apply 3 field changes to the survivor".
For Account, Contact and Lead you choose a strategy:
- Native merge: Salesforce reparents the losers' child records to the survivor, then deletes the losers.
- Keep master + delete losers: no reparenting; the losers go to the Recycle Bin.
Other objects get only the second strategy, with a warning that child records are not moved.
Type MERGE, then click Merge records (native merge) or Merge & delete losers.
What it reads and writes
The scan is one SOQL query for the record Id and your match fields, limited to 2,000 records. The comparison runs in your browser. Resolving re-reads the group's records, up to 140 editable fields.
- Native merge uses the SOAP API
mergecall, which takes the survivor and up to two losers. Larger groups take several calls in a row, and the calls stop at the first failure. Your field choices are sent as values on the survivor. - Keep master + delete losers first updates the survivor with your choices through the REST API. If that update fails, nothing is deleted. Then it deletes the losers.
The extension has no undo for either strategy. Losers deleted by the second strategy can be restored from the Recycle Bin.
Limits
- A scan reads at most 2,000 records, in no set order. Past that, the results say capped at 2000 and duplicates among the other records are not found.
- The resolve window compares and copies at most 140 editable fields. Address, geolocation, encrypted and binary fields are left out.
- With Native merge, picking (blank) for a conflicting field does not clear it; the survivor keeps its own value.
- You resolve one group at a time. Resolved groups are hidden until you scan again.
Related
- Data Quality: score an object for missing values and bad email, phone and URL formats
- Related Records: see what is connected to a record before you merge or delete it
- SOQL Builder: query the records in a group in more detail