Use matching signals carefully, retain separate enquiries and verify the history and ownership that survive a CRM contact merge.
A customer calls the rental desk, sends a message later and then appears again in a website enquiry. Three entries may represent one person. They may also represent two colleagues using the same office number. Treating every match as a duplicate can be as damaging as keeping every entry separate.
Duplicate-contact cleanup is an identity decision. It should preserve the evidence behind the decision and the work attached to the records. For a growing Nepal business, a careful merge process is more useful than a quick reduction in the number of rows.
Separate a person from an enquiry
A contact represents an identified person or business relationship. An enquiry represents a particular request. The same person can have several legitimate requests: hiring a generator for one event and booking lighting for another is not necessarily duplicate demand.
A hypothetical Kathmandu equipment-rental business receives messages from a coordinator using a personal mobile and a shared company email. Staff should link confirmed identities while retaining the separate event requirements, dates and quotations. Merging contact information must not quietly combine those jobs.
Illustrative planning scene with conceptual software screens; not real staff or client evidence.
Use matching signals as evidence rather than proof
Start with the fields your team actually collects: stable internal ID, verified email, phone, company and supporting conversation history. Standardise formatting carefully. Removing spaces in a phone number can help comparison; deleting meaningful digits or guessing a country code can create a false match.
HubSpot’s official deduplication documentation describes matching contacts through email and using record IDs or unique properties in imports. That vendor-specific behaviour demonstrates why identifiers matter. It is not a universal rule that an email address identifies one person, nor evidence of WorkSphere’s matching capabilities.
| Signal |
Useful question |
| Internal record ID |
Is this the existing retained identity? |
| Email or mobile |
Verified personal detail or shared contact? |
| Company and name |
Corroborating evidence or only similarity? |
| Linked enquiries |
Separate legitimate jobs to retain? |
Treat a shared switchboard number, common surname or similar company name as a reason to investigate. Agree which evidence allows a reviewer to confirm a match. When evidence conflicts, keep the records separate and flag the uncertainty instead of forcing a cleaner-looking list.
Suggested decision flow. Review the actual software behaviour and approval rules before implementation.
Count removals without overstating what you know
Suppose a hypothetical contact list contains thirty records. Reviewers confirm four separate duplicate pairs, with no record appearing in more than one pair. Merging each pair into one retained record removes four redundant records: thirty minus four equals twenty-six retained contacts.
HYPOTHETICAL / four disjoint duplicate pairs. Candidates are not confirmed duplicates. Retained records are not a verified count of people.
Six additional records form three candidate pairs that still need review. They remain in the twenty-six until a decision is made. Do not subtract all ten pairs or claim the list now represents twenty-six verified people. The example concerns records and decisions, not proven customer identity or sales results.
Hypothetical workflow dashboard; not a live WorkSphere screenshot. Pair counts, record counts and enquiry counts are different units.
A useful cleanup panel separates confirmed merges, unresolved candidates and rejected matches. Its counts should be mutually exclusive. A rejected match may be correct evidence that two similar records describe different people, rather than a failed cleanup task.
Decide what survives before merging
Name the retained record and identify the value to keep for each conflicting field. A recently typed phone number is not automatically more reliable than an older verified one. Preserve useful alternative contact details with their context instead of overwriting them without explanation.
Check linked quotations, activities, consent records where applicable, open tasks and customer-account references. Keep a before-and-after record of the decision. If an accounting customer identity is involved, the accounts owner should assess the effect before the CRM reviewer proceeds.
Conceptual illustration of the underlying decision; no measured outcomes or exact figures represented.
Do not transfer every permission or communication preference by assumption. A merge must not broaden who receives messages or who can see confidential history. Ask the implementer how linked records are handled and what can be reversed if the match later proves wrong.
Test one merge and inspect the linked work
Use an anonymised test pair in a safe environment before cleaning a large list. Confirm that both records’ relevant history can still be found and that separate enquiries remain separate. Check which owner receives the open tasks and whether reports now count the retained contact consistently.
Illustrative review scene; any screen shown is conceptual, not a live product screenshot.
Hypothetical identities. No customer data or live CRM behaviour is represented.
Assign a reviewer for uncertain matches and keep a short decision note. The person who spotted the duplicate may provide evidence, while someone familiar with the customer confirms it. That division is a suggested routine; small teams can adapt it without pretending every match requires a committee.
Prevent the same problem from returning
Search existing records before creating another contact. Ask import owners to provide the approved identifier when updating records. Define how website messages, phone enquiries and other sources look up an existing person without assuming that every channel supplies a reliable unique key.
A CRM can support these controls, but the exact lookup, merge and audit behaviour needs verification. For a proposed WorkSphere workflow, request a demonstration using a shared phone number, a changed email and two separate requests from the same person. Those examples reveal more than a promise of “automatic deduplication”.
HubSpot Deduplicate Records