Skip to main content

Potential Duplicates Component

Explore the Potential Duplicates lightning component that surfaces possible duplicate records directly on record pages in real time. Learn how to add the component to a page layout, review and compare potential matches, merge or dismiss duplicates, and customize display fields and detection thresholds.

What You'll Learn

  • Add the Potential Duplicates lightning component to any standard or custom record page
  • Pick up to five matching rules per object via the gear icon configuration
  • Configure cross-object matching to surface Lead/Contact duplicates
  • Merge or dismiss duplicates directly from the record page without navigating away
  • Understand why component results may differ from scheduled report results

Key Takeaways

  • The component evaluates duplicates in real time when the record page loads
  • Matching rule selection is per object, not per Lightning page
  • Read-only users see duplicates but cannot merge — by design
  • Limit to focused, fast matching rules to keep page load times responsive

Frequently Asked Questions

How many matching rules can the Potential Duplicates component use?

Up to 5 matching rules per object, with at most 1 cross-object matching rule. Selection is saved per object, so configuring rules for Contact applies to every Contact record page where the component is placed.

Why are the results different from the scheduled duplicate report?

The component evaluates rules in real time against the current record data, while scheduled reports may use cached results from the last scan. Different rules may also be selected in each context.

Can read-only users merge duplicates from the component?

No. Users with the No Duplicates Read Only permission set see the component and the duplicate list, but the Merge button and the gear icon are hidden. Only users with the admin permission set can merge or change rule selection.

Does the component slow down record page loading?

It can if you select many matching rules or rules that scan very large datasets. Limit to focused, fast rules — typically 2 to 3 — for good page performance, especially on high-traffic objects.