Glossary term
Data Deduplication
Data deduplication is finding records that represent the same person or company more than once and merging them into a single record that keeps the most accurate values and all of the history attached to both.
What is data deduplication?
Data deduplication in a CRM is the process of finding multiple records that point to one real entity, such as a contact entered twice or a company saved under two names, and consolidating them into one surviving record. It runs in three steps. Matching identifies suspected pairs using keys such as email address, mobile number or website domain. Survivorship decides which value to keep for each field where the two records disagree. Merging moves everything attached to the duplicate (deals, messages, notes) onto the survivor. Duplicates do more than distort reports; they cause mistakes the customer notices: two reps calling on the same day, a campaign arriving twice, or a relationship history split across two records so nobody ever sees it whole.
Why duplicates happen
| Cause | Example |
|---|---|
| Names written in two scripts or spellings | The same person saved once with an Arabic spelling and once as 'Mohammed Alotaibi' |
| Phone number formats | One number stored in local format starting 05 and in international format starting +966 |
| Imports from several spreadsheets | A trade-show visitor list imported on top of an existing database |
| Several channels | A buyer fills in the website form, then messages on WhatsApp from a different number |
| Company naming | 'Horizon Trading Co.' and 'Horizon Trading' sharing one email domain |
Matching keys differ in strength. Email address and company domain are strong; a mobile number is strong once its format is normalized; a name on its own is weak, especially for common names, and should never drive a merge without human review.
Deduplication also supports data accuracy under Saudi Arabia's Personal Data Protection Law. SDAIA's guide for controllers says personal data should be reviewed regularly to keep it accurate, complete and up to date, and that individuals should be able to review and update it.
In Naffes CRM
Duplicate detection in Naffes CRM runs as a background scan. Contacts are matched by email, phone or name, and companies by domain or name. Suggested pairs are then reviewed in the Data Quality Center before anything is merged, so no two records are combined without a person deciding.
Start the scan
On the Contacts page, select records and choose Merge duplicates in the bulk actions bar.
Review the pairs
When the scan finishes, open Data Quality and then Dedup, or use the Merge duplicates button on the Contacts toolbar.
Pick values and merge
For each pair, choose the value to keep in every field and click Merge, or dismiss the pair if it is not a duplicate.
A merge moves all related records to the surviving record and soft-deletes the other. The data health dashboard completes the picture with an overall score and a list of gaps: contacts without email, phone or owner; companies without domain or owner; and open deals with no stakeholders, no activity for 30 days or more, or no expected close date.
- Known numbers link up: a WhatsApp or SMS message from a number already in the CRM attaches to the existing contact instead of creating a new one.
- Active-deal warning: a contact's full page shows a banner when the contact already has an active deal, even one the rep cannot open, which heads off duplicate deals.
- Unique fields: contacts and companies support unique custom fields, such as a commercial registration number, so the same value cannot sit on two records.
Plan and scope: Data Quality, including deduplication, is a Pro and Enterprise feature. The scan covers contacts and companies, not deals; a company scan cannot be started from the UI yet, and merges cannot be undone from the UI.
Habits that stop duplicates before they start
- Normalize phone and email formats in a file before importing it, and map fields carefully in the import wizard.
- Scan for duplicates after every large import, not once a year.
- Agree on one way to write company names, and on one unique field that identifies each company.
- Move customer chats onto a channel connected to the CRM, where a known number links to its owner automatically.
Clean data underpins everything built on it: lead scoring loses meaning when a buyer's activity is split across two records, the activity timeline tells half the story, and a sales forecast inflates with duplicate deals.
Questions
Questions 2
Do I lose deals or messages when two records are merged?
No. In Naffes a merge moves everything linked to the duplicate onto the surviving record, and you choose which value stays in each conflicting field. Merges cannot be undone from the UI, though, so check each pair before confirming.
Will Naffes match an Arabic name with its English spelling?
The scan matches contacts by email, phone or name, so two records written in different scripts are usually caught through a shared email or phone number. That is one more reason to keep a normalized email or mobile number on every contact.