Salesforce duplicate rules that protect pipeline without killing conversion
Most duplicate setups fail because one policy is forced across every channel. The fix is operational: match strictness by source, queue ownership for merge decisions, and weekly tuning based on duplicate-rate drift.
Duplicate management is a routing problem, not a checkbox
Treat this as pipeline hygiene with owners and service levels.
policy layers (website, manual, import)
queue owner for merge decisions
days between duplicate-rate reviews
Set duplicate actions by channel, not by object alone
Website intake should optimize conversion. Imports and back-office edits should optimize strictness.
Website Leads
Default to allow + alert. Route likely duplicates to review queue without blocking form submissions.
Manual Entry
Use warning-first enforcement for SDR/admin users. Require explicit override comments for risky saves.
Bulk Import
Pre-dedupe before upload, then run post-import duplicate reports with mandatory cleanup window.
Three policy calls that prevent duplicate-rule regret
Block all website duplicates by default
Allow + alert for website leads, then triage in queue
Match only on email
Use email + domain + normalized phone for lead matching
Treat merge as occasional cleanup
Run daily merge operations with an owner and SLA
A practical duplicate-control loop
Define policy
Set allow/warn/block per channel and object risk profile.
Tune matching
Adjust match strictness with real inbound samples.
Operate merge queue
Run daily triage with attribution-safe merge checklist.
Review weekly
Track duplicate rates by source and spot drift early.
Identity field specification by object
Document this before tuning matching rules. If identity keys are ambiguous, duplicate behavior will drift every quarter.
Lead object
- Primary: lowercased email + normalized company domain
- Secondary: E.164 phone + last name
- Fallback: first name + last name + company
Contact object
- Primary: email + account ID
- Secondary: phone + account ID
- Fallback: full name + account domain
Account object
- Primary: canonical domain
- Secondary: normalized legal name
- Fallback: phone + billing country
Test scenarios before production rollout
Run these as scripted submissions so policy changes are validated against real lead patterns, not assumptions.
Scenario A: net-new website lead
Expected: save succeeds, no duplicate alert, normal routing.
Scenario B: same person, slightly different company name
Expected: allow + alert for website path, record enters duplicate review queue.
Scenario C: bulk import of known contacts
Expected: high warning rate surfaced in post-import duplicate report and merge backlog created.
Need help fixing duplicate rules without sacrificing lead volume?
We redesign duplicate policy, tune matching logic, and stand up merge operations your RevOps team can actually sustain.