Get rid of duplicate customer records for good
John Smith. J Smith. John Smith Ltd. Jon Smith with the typo. Somewhere in there is one actual customer, spread across four records.
Where it hurts
Duplicates creep in quietly. Someone adds a customer who's already there. An import runs twice. A name gets spelt two ways. Before long your list is a third bigger than your customer base, and each version holds a different piece of the truth.
It makes everything harder. You send two statements, credit the wrong record, and can't trust any total because the customer's history is split across copies. Reports lie to you.
Cleaning it by hand is grim work, so it never gets done, and the mess just grows.
What I'd build
I'd start with a proper de-duplication pass across your CRM, accounts package, or both: matching on more than just name (phone, email, address, so 'J Smith' and 'John Smith' get spotted as the same person), then merging carefully so nothing is thrown away. The risky merges happen with you watching, not blind. Then I'd stop it happening again: the systems get connected so a change in one place lands in the others, and an AI check watches new records as they're created, flagging likely duplicates and odd entries for a human to confirm before the mess rebuilds itself.
What changes
- One clean record per customer, with every note and invoice kept
- Systems connected so details stay in step from now on
- AI flags new duplicates before they take root
- Risky merges happen with you watching, not blind
Related fixes
Sound like your week? The free half-hour review is where we work out whether this is your first fix or your fifth.