Stop Reopening Duplicate Contact Reviews Without New Evidence
Prepared with AI assistance by Alactic Editorial. Examples identified as hypothetical are illustrative, not customer results.
A contact pair returns for review because the names match. Someone checks it and leaves the records separate. Later, another colleague notices the same match and starts again. The problem may not be poor judgment: the earlier reasoning is missing from the place where the next decision happens.
To stop repeated false duplicate contact reviews, record a durable, pair-specific decision: which records were compared, why they remain separate, what evidence was considered, and what would justify reconsideration. A new reviewer, a later date, or another appearance of the same matching field is not new evidence.
A keep-separate register makes that decision discoverable without making it immune to challenge. The goal is to prevent repeated guesswork, not to preserve a mistaken conclusion forever.
1. Distinguish supported separation from unresolved identity
Closing a review does not necessarily settle whether two records represent different people. Sometimes the evidence supports separation; sometimes it merely fails to justify combining them.
Use two explicit outcomes:
- Keep separate—supported: the evidence supports treating the records as different identities. The rationale explains what distinguishes the people, not just their contact details.
- Keep separate—unresolved: the evidence does not justify combining the records, but it also does not establish that they represent different people.
Both outcomes leave the records separate. Their evidential meaning differs. Two accurate addresses, business roles, or correspondence histories can belong to one person. Confirming those details does not, by itself, establish two identities.
An unresolved decision is therefore useful, not incomplete paperwork. It states the current boundary of knowledge and identifies what could resolve it. Do not upgrade uncertainty to supported separation merely because staff have reviewed the pair several times.
Keep the register focused on identity decisions. It does not certify every contact field, configure duplicate detection, or replace general follow-up management.
2. Make the decision belong to the pair
A note saying “not a duplicate” on one contact leaves questions unanswered: not a duplicate of which record, based on what evidence, and under what circumstances?
Identify both records with stable references alongside their display names. Names, phone numbers, and email addresses help reviewers recognise a pair, but can change or be shared.
Each register entry should contain:
- Pair reference: a unique label for the comparison.
- Record references: both contact identifiers and display names.
- Decision: supported or unresolved separation, with reviewer and date.
- Trigger: the similarities that prompted comparison.
- Evidence: retrievable source references and what each establishes.
- Rationale: why the evidence supports this outcome, including its limits.
- Reopening condition: evidence or a reasoning defect that could change the assessment.
- Non-trigger: what does not justify another full review.
- History: challenges, dispositions, corrections, and superseding decisions.
Use a shared location accessible to the appropriate reviewers. Where your workspace supports it, make the entry discoverable from either contact. Otherwise, use a consistent reference convention. Check the workspace and current plan before relying on any particular feature; this is an operating practice, not an assertion about an Alactic control.
Retain enough evidence to check the reasoning without copying entire correspondence archives. A precise source reference and factual summary are usually easier to review than an unexplained collection of screenshots.
3. Reusable artifact: a filled keep-separate register
Fictional worked example: All names, identifiers, dates, and evidence in this register are hypothetical. The entries illustrate a recommended design, not Alactic customer results. Reuse the field structure, replacing the example material with evidence from your own review.
KS-101: Namesakes with separate histories
- Records: C-104, Jordan Lee, North Pier Supplies; C-882, Jordan Lee, Willow Studio.
- Decision: Keep separate—unresolved. Priya Shah, 12 May, hypothetical review date.
- Trigger: identical full names and the same town.
- Evidence: correspondence references NP-12 and WS-08 show different business roles and contact addresses. Each correspondent confirms their own record details.
- Rationale: the details may be accurate, but one person could hold both roles and use both addresses. Neither the matching name nor the separate histories establishes identity. There is insufficient evidence to combine the records or confidently declare two people.
- Reopen if: attributable clarification connects both histories to one person or directly distinguishes the two people, explaining the relationship between the records.
- Non-trigger: the same name-and-town match, another copy of the import, or reconfirmation of an address without identity clarification.
- History: initial unresolved decision; no claim of proven separate identities.
KS-102: Two people using a household number
- Records: C-215, Amira Patel; C-216, Dev Patel.
- Decision: Keep separate—supported. Mateo Ruiz, 13 May, hypothetical review date.
- Trigger: matching surname, address, and telephone number.
- Evidence: correspondence references HP-21 and HP-22 explicitly identify Amira and Dev as two different household members sharing the telephone. Each distinguishes their own service history from the other's.
- Rationale: the evidence directly describes two people and attributes their separate histories. The shared number is a contact route, not a unique personal identifier.
- Reopen if: credible source evidence shows a name or service history was attributed to the wrong person.
- Non-trigger: another message from the household number or another match on the address.
- History: initial decision; shared-number explanation recorded.
KS-103: Different staff behind a business number
- Records: C-330, Elena Torres; C-471, Marcus Chen.
- Decision: Keep separate—supported. Priya Shah, 14 May, hypothetical review date.
- Trigger: matching reception number and company mailbox.
- Evidence: correspondence reference BC-14 identifies Elena and Marcus as separate employees, assigns their respective purchasing and service responsibilities, and confirms that both use the shared channels.
- Rationale: the source distinguishes two people; it does not merely list different roles. Shared organisation details explain the apparent match.
- Reopen if: a reliable correction establishes that either named entry or its attributed history is wrong.
- Non-trigger: matching company details or a different employee answering the shared line.
- History: initial decision; no superseding review.
KS-104: Disputed match with insufficient evidence
- Records: C-509, Sam Okoro; C-938, Samuel Okoro.
- Decision: Keep separate—unresolved. Mateo Ruiz, 15 May, hypothetical review date.
- Trigger: similar names and the same email address.
- Evidence: import IM-09 contains the matching email, but its original attribution is unknown. One staff note calls the records duplicates; another disputes that conclusion. Neither cites supporting material.
- Rationale: combining the records would treat uncertain attribution as proof. The opposing assertions also do not establish separate identities.
- Reopen if: the original source establishes who supplied the address, or attributable clarification connects or distinguishes the histories.
- Non-trigger: the same imported email, another unsupported assertion, or elapsed time alone.
- History: unresolved separation; awaiting evidence rather than repeated opinion.
4. Test whether proposed evidence could change the answer
“Review if anything changes” is too broad. A corrected postcode can be new without answering the identity question. “Never merge” is too rigid because it prevents correction of demonstrated errors.
Assess a challenge using three questions:
- Is it newly considered? Another copy of the original import is not independent corroboration. An older source overlooked in the first review can still be newly considered evidence.
- Is it relevant? Explain how it addresses the recorded uncertainty or undermines the rationale.
- Is it sufficiently attributable? Can reviewers check its origin and connection to the records concerned?
The threshold is enough reason to reconsider, not enough proof to reverse the decision immediately. A credible contradiction may warrant investigation before its full implications are known.
A demonstrated reasoning defect can also justify reopening without new identity evidence. For example, if KS-101 had been labelled supported solely because the addresses differed, a challenge could correctly identify that one person may use both. Reconsidering that label repairs an overstatement; it does not prove the records should be combined.
5. Separate a challenge from a reopened review
Allow colleagues to challenge a decision without automatically restarting the entire comparison. Ask for the pair reference, the disputed rationale, and the proposed evidence or reasoning defect.
If the challenge does not warrant reopening, record the actual reason, rather than labelling every unsuccessful submission repeated evidence:
- Repeated evidence: the submission adds no material beyond the sources already considered.
- Irrelevant change: the information is new but does not address the identity question.
- Insufficiently attributable evidence: the information could matter, but its source or connection to the pair cannot yet be checked.
- No demonstrated reasoning defect: the challenge questions the conclusion without showing an error in the recorded reasoning.
When a challenge is potentially relevant, state what clarification would make it actionable. This might be the original message, identification of which records the sender meant, or an explanation connecting a changed detail to the identity question. A declined challenge can remain open to clarification without becoming a full reopened review.
Worked continuation of KS-101, entirely hypothetical: a colleague submits a forwarded note saying that Jordan changed businesses. That could explain the two histories, but the note does not identify its author or either contact record. Record: “Not reopened: insufficiently attributable evidence. Please provide the original correspondence and clarify whether it refers to both C-104 and C-882.” Do not call it repeated evidence.
If attributable correspondence subsequently connects both histories, open a linked review. Record what became checkable, who assessed it, and the resulting decision. The reopened review may still end in separation: reopening is justified by a meaningful question, not by a requirement to reverse the result.
6. Repair weak decisions without erasing history
A register can fail through missing reasoning, excessive certainty, or poor discoverability. Repair the specific weakness:
- The old note says only “not a duplicate.” Treat it as a historical assertion. Reconstruct the rationale from available sources; if that is impossible, mark separation unresolved.
- A source is no longer accessible. Record the gap and assess whether remaining evidence supports the conclusion. Do not imply the missing source was checked again.
- The rationale overstates the evidence. Correct the label and explanation. Preserve the earlier version and state why it changed.
- A third contact appears. Review the relevant new pairs. A decision about two records does not settle every relationship involving another record.
- Reviewers cannot find the entry. Improve the shared reference convention and visible summary rather than creating another private register.
Preserving history does not mean preserving error. It means making the correction understandable to the next reviewer.
7. Check whether the register reduces repeated guesswork
Start with recurring pairs from actual review records. Distinguish repeat submissions that merely reproduce the original match from challenges containing relevant evidence or a reasoning defect.
Hypothetical starting exercise: examine ten recently repeated pairs. For each, record whether a prior rationale existed, whether reviewers could find it, and whether the latest submission added anything material. Create supported or unresolved entries according to the evidence, not a target proportion of resolved cases.
At the next agreed checkpoint, inspect each return of those pairs. Can reviewers close unchanged submissions by referencing a clear rationale? Are declined challenges given accurate reasons? Do potentially relevant but unattributed submissions receive a useful clarification request?
A lower reopening count alone is not success: staff may have stopped raising legitimate concerns. The observable improvement is more specific—unchanged matches no longer trigger full reconstruction, while meaningful challenges receive documented consideration and errors remain correctable.
Put the next conversation to work.
Explore Alactic Communications for WhatsApp conversations, campaigns and AI-assisted workflows. Choose the product and plan that fit your team.
Explore Communications