Finding and Resolving Duplicate Profiles
Duplicate profiles creep in over time: someone registers twice with different email addresses, a CSV import overlaps with people who already signed up through a form, or a staff member creates "Bob Smith" when "Robert Smith" already exists. The Find Duplicate Profiles tool scans every profile in your organisation, shows you pairs that are likely to be the same person with a confidence score, and lets you merge them or mark them as different people.
This article covers how the scan works, how to review its results, and exactly what happens to a person's data when two profiles are merged. For merging a handful of profiles you have already found yourself, see Merging Duplicate Profiles.
Opening the duplicate finder
From the dashboard, open People and choose Find Duplicate Profiles. The same tool is also listed under Utilities on the system dashboard.
The first time you visit, you'll see No scan yet. Click Scan for duplicates to start.
Running a scan
- The scan runs in the background. A progress bar shows the percentage complete, and you can leave the page and come back — the latest scan is remembered.
- The scan covers every profile in your organisation, including inactive ones. Deleted profiles (in the trash) are not included.
- Once finished, the page header shows when the results were produced (for example "Scanned 3 days ago").
- Results don't refresh by themselves. When you've imported new people or made lots of changes, click Rescan to run a fresh scan. A rescan replaces the results on screen.
- If a scan fails, you'll see Scan failed with a Retry scan button.
For very large organisations, a few unusually large clusters of similar records may not be fully compared. When that happens a notice reads "Some very large groups couldn't be fully compared." The rest of the results are still valid.
How matching works
Comparing every profile with every other profile would be slow, so the scan first sorts people into small "candidate" buckets of profiles that share something — a similar-sounding surname with the same first initial, the same full name, the same email username, the last digits of a phone number, a surname plus birth year, or a surname plus postcode. Only profiles that land in a bucket together are compared in detail.
Names, nicknames and misspellings
- Nicknames — a built-in list of common nickname pairs means "Bob" and "Robert", or "Jess" and "Jessica", are treated as a name match.
- Sound-alike surnames — surnames are also compared by how they sound, so "Thomson" and "Thompson" can still be found.
- Typos — first and last names that are very similar (one or two letters different) count as a partial match.
- Alternative names — any alternative first or last names stored on a profile (for example from a previous merge) are compared too.
- Combined names — if a full name was entered entirely in the first name field ("Jane Smith" with no last name), the scan works out the surname for comparison purposes.
Other details compared
| Detail | How it counts |
|---|---|
| Phone number | An identical phone number is one of the strongest signals. |
| Identical emails score highly; very similar emails, or the same email username at a different domain (jane.smith@example.com vs jane.smith@example.net), also count. | |
| Address | Identical or similar street address, same suburb, same postcode. |
| Date of birth | Exact match, partial match, or same birth year. |
| Gender | Same gender adds a little; a different gender counts strongly against a match. |
Signals that lower confidence
The scan also looks for the patterns that usually mean family members rather than duplicates, and reduces the score accordingly:
- Age gap — dates of birth more than five years apart (a gap of 15+ years suggests parent and child).
- Household sharing — the same phone, email and/or address but clearly different first names, which is typical of partners or a parent and child sharing contact details. A child's contact details are treated as likely belonging to a parent.
- Likely family — the same last name with very different first names is capped at a low confidence.
- Senior/Junior — one name marked Senior/Snr/Jr and the other not.
Pairs scoring 55% or higher are shown as possible duplicates. Pairs you've previously marked as distinct are never shown again.
Reviewing the results
Results are listed highest confidence first, 25 at a time (use Load more for the next page). Each card shows:
- A coloured dot and a summary such as "We're 92% confident these 2 profiles are duplicates".
- Badges explaining why — for example Identical phone, Similar email (90%), Nickname Match, Same postcode, Identical DOB. Warning badges such as Gender mismatch, 12yr age gap, Household sharing, Likely family or Senior/Junior mismatch flag reasons to be careful.
- Each profile's name, gender, age, emails, phone numbers and address, plus its status. Click View to open the full profile, or the … button for the usual profile actions.
Merging a pair
- Check the two profiles really are the same person — open each with View if you're unsure.
- Click Merge. For pairs below 75% confidence you'll be asked to confirm.
- A progress bar shows the merge running. When it finishes the card reads Merge complete!.
- Click View Merged to check the combined profile, then Dismiss to remove the card.
Marking a pair as different people
Click Mark Distinct when two profiles belong to different people (twins, a parent and child with the same name, and so on). The pair is remembered permanently and won't appear in this scan or any future scan.
Merge All 100%
Once a scan is complete, Merge All 100% merges every pair the scan is 100% confident about — across the whole result set, not just the cards currently loaded. Pairs are merged one after another, the header shows "Merging 3 of 40", and you can click Cancel to stop after the current pair. Any pair that couldn't be merged stays in the list. Merged pairs don't return in later scans.
Merging is hard to undo. Skim the 100% pairs before using Merge All, especially if your data contains families who share an email address or phone number.
What happens when profiles are merged
Whether you merge from the duplicate finder or from the Merge action on the people list, the same rules apply.
Which profile survives
The profile that was created first is kept. The other profile is moved to the trash, and its activity is recorded on both profiles' history.
How fields are combined
| Field | Result |
|---|---|
| First and last name | The surviving profile's names are kept. Every other first and last name is saved as an alternative first name / alternative last name, so searches for the old name still find the person. |
| Emails, phone numbers, addresses and other multi-value fields | All values are combined and duplicates removed. Phone numbers and addresses written slightly differently are treated as the same entry. |
| Gender | The most detailed value is used (a known gender wins over "unknown"). |
| Status | If either profile is active, the merged profile is active. |
| Yes/no fields | If either profile says yes, the result is yes. |
| Date fields | The most recent date is used. |
| Other single-value fields | The value from the most recently created profile that has one is used. Blank values never overwrite filled ones. |
| Created date | The earliest created date is kept. |
| Read-only fields | Not changed. |
When you merge from the people list, the preview page lists every field that differs, what each profile holds, the action to be taken and the result — check it before clicking Start Merge. Date of birth is stored as separate day, month and year values, so look closely at it when the two profiles disagree.
What moves to the surviving profile
- Everything that points at the person — check-ins, form submissions, workflow cards, roster assignments, conversations and any other record that references the merged-away profile — is updated to point at the surviving profile. This happens in the background; the progress bar tracks it.
- Group and team positions — memberships move across. If both profiles held the same position, it is kept once.
- Scopes — the surviving profile ends up in every scope either profile belonged to.
- Logins — any login linked to the merged-away profile is linked to the surviving profile, so the person can still sign in and sees their combined history.
- Comment and message read status and conversation mute settings follow the person.
Permissions
- To merge, you need permission to edit the surviving profile and to delete every profile being merged into it (admins can always merge).
- Marking a pair as distinct records the decision on both profiles.
- The scan itself looks at every profile in the organisation, so give access to this tool only to people who should see your whole database.
Tips
- Run a scan after every large import — it's the quickest way to catch overlaps with existing records.
- Work from the top: high-confidence pairs usually take seconds; slow down on anything with a warning badge.
- Use Mark Distinct generously for genuine look-alikes so they stop cluttering future scans.
- Deleting a profile removes any open duplicate cards that include it. Restoring it later doesn't bring them back — run a fresh scan.