Salesforce data quality · Duplicates

Find, merge, and un-merge duplicate records in Salesforce

Duplicate Leads, Contacts and Accounts inflate your pipeline, split activity history, and route the wrong rep to the wrong deal. Here’s how to find duplicate clusters, merge them with golden-record matching, and — the part that’s usually missing — undo a merge when a match was wrong.

See your duplicate clusters (free scan) View the interactive preview
The problem

Duplicates don’t look like a problem — until the forecast does

Two Accounts for the same company means coverage counted twice, activity split across records, and a rep working a relationship someone else already owns. Multiply that across a year of imports and hand-keyed records, and the pipeline you’re forecasting on is quietly inflated. For the broader Salesforce data quality remediation lifecycle, see the complete playbook.

Native tools vs. deeper remediation

Salesforce Matching Rules and Duplicate Rules provide a useful foundation

Salesforce ships Matching Rules and Duplicate Rules that can flag potential duplicates at the point of entry and surface merge suggestions in the UI. For teams running standard processes at manageable record volumes, these built-in capabilities provide real coverage. Where organizations typically seek additional capability: higher-volume environments where the manual merge backlog accumulates faster than admins can clear it, legacy data that pre-dates the current rule set, and situations that call for cluster-level golden-record identification — grouping every variant of the same real-world entity rather than comparing pairs — so that a single authoritative record can be confirmed and preserved across the full set. ForecastGuard is built for teams who need that second layer.

How to do it well

Golden-record matching, across the whole cluster

Instead of comparing two records at a time, golden-record matching clusters every variant of the same real-world entity and identifies the single surviving record — the most complete, most current version — across the set. That’s master-data logic, run natively against Sales Cloud rather than in a separate platform.

STEP 01

Cluster the duplicates

Group every variant of the same Lead, Contact or Account — not just exact matches — into one cluster.

STEP 02

Confirm the golden record

Identify the single surviving record to keep, built from the most complete and current values in the cluster.

STEP 03

Merge — reversibly

Remediate the cluster in one click, with a steward’s undo if a match was ever wrong.

The part that’s usually missing

Merge in one click — with a steward’s undo

The reason teams hesitate to merge is the fear of losing data to a wrong match. ForecastGuard makes remediation reversible: merge a cluster in one click, and if a match was wrong, unmerge that single cluster — or roll back every merge at once. It’s remediation you’ll actually run, because you can reverse it. And it isn’t only Accounts: ForecastGuard clusters and remediates across all five core objects — Leads, Contacts, Accounts, Opportunities and Cases — without exporting data or standing up a separate system.

See your own data

Find out how dirty your Salesforce data really is.

Start free. Get the 2026 Dirty Data Index and a scan of your org, or score your forecast risk in two minutes — no email required.

Get the Dirty Data Index scanTake the Forecast Risk Assessment
Questions buyers ask

Salesforce duplicate management, answered

What is Salesforce duplicate management?

It is the process of finding duplicate Leads, Contacts, and Accounts, identifying a golden record, and merging the cluster.

How does golden-record matching differ from pairwise matching?

It groups every variant of the same real-world entity into one cluster instead of comparing only two records at a time.

Can a Salesforce duplicate merge be reversed?

ForecastGuard makes remediation reversible: a steward can unmerge a single cluster or roll back every merge.

Does ForecastGuard cover more than Accounts?

Yes. It clusters and remediates Leads, Contacts, Accounts, Opportunities, and Cases without exporting data or standing up a separate system.