Duplicate records in your CRM: prevent and merge them

Duplicate records are the quiet tax on every CRM. Two reps call the same homeowner. A customer gets the same text twice. Your lead source report counts one buyer as three leads. This guide explains why duplicates happen, how to match records reliably, how to stop them at the source and how to merge the ones that slip through without losing history.

Why duplicates matter

Duplicates cause problems well beyond a messy contact list:

• Awkward customer moments. Two reps follow up with the same lead, or a customer gets a “welcome” message months after they signed.
• Broken history. Half the conversation sits on one record and half on another, so nobody sees the full picture.
• Wrong numbers. Lead counts, conversion rates and source reports are inflated.
• Compliance risk. If someone opts out on one record but a duplicate still allows messages, you may contact a person who asked you to stop.

The problem is common. In Salesforce’s 2026 State of Sales survey of 4,050 sales professionals in 22 countries, duplicate data ranked as the second most common data issue among sales professionals on teams using AI agents, behind manual errors, and 46% of those with agents said data quality issues hurt their sales (Salesforce State of Sales, 7th edition). That’s vendor research from a CRM company, but it matches what most small teams see in practice.

Why duplicates happen

Most duplicates come from the same person reaching you through more than one door:

• Multiple entry points. A homeowner fills in your web form, then calls, then texts from a different number.
• Imports. A list from a trade show or an old spreadsheet is imported without checking against existing records.
• Manual entry. A rep adds “Jon Smith” without searching for “Jonathan Smith.”
• Formatting differences. “(512) 555-0142” and “+15125550142” look different to a system that doesn’t normalize them.
• Several tools. Data copied between a CRM, an accounting tool and a texting app drifts out of sync. In a 2024 Salesforce survey of 3,350 leaders at businesses with 200 or fewer employees, 53% agreed that data is often inconsistent across their business tools (Salesforce SMB Trends, 6th edition).
• Resubmissions. People submit the same form twice when they don’t get a quick reply.

How to match records

Good duplicate detection compares the fields that identify a person, in a standard format, and treats weaker matches with care.

Email

Email is the strongest identifier for most leads. Before comparing, trim spaces and convert to lowercase. Strictly speaking, the email standard, RFC 5321, says the part before the @ must be treated as case sensitive, but it also says exploiting that “impedes interoperability and is discouraged,” and domain names are not case sensitive. In practice, treating emails as case-insensitive is the sensible choice for a CRM.

Watch out for shared or role addresses like info@ or office@. Several people at the same company may use them, so a match on a role address shouldn’t merge two different contacts automatically.

Phone

Phone numbers need normalizing before they can be compared. Store them in one international format: a plus sign, the country code and the number, based on the ITU’s E.164 numbering plan. Then “(512) 555-0142,” “512.555.0142” and “+1 512 555 0142” all become the same value.

Remember that families and small offices share numbers, so a phone match is strong but not always conclusive.

Name and company

Names are the weakest signal. “Maria Garcia” might be two different people; “Bob Lee” and “Robert Lee” might be one. Use name matches, ideally combined with company or address, as suggestions for a person to review, never as a reason to merge automatically.

Match confidence

Match	Confidence	What to do
Same normalized email (personal address)	High	Flag as duplicate; merge after a quick check
Same normalized phone and similar name	High	Flag as duplicate; merge after a quick check
Same phone, different names	Medium	Review; could be a household or shared office line
Same role email (info@, sales@)	Medium	Review; may be different people at one company
Similar name and same company or address	Medium	Review
Similar name only	Low	Ignore unless other details line up

Prevent duplicates at entry

The cheapest duplicate is the one that never gets created.

• Live checks while typing. When someone adds a lead or contact, the CRM should warn them if the email, phone or name already exists.
• Search before you create. Make it a team habit, especially for phone leads.
• Update, don’t create, on known contacts. When an existing customer fills in a form or texts in, the activity should land on their existing record.
• Use required formats. Phone fields that store numbers in one format make matching reliable.
• Keep forms short and clear. Fewer fields means fewer typos. Our guide to lead capture forms covers form design.

Prevent duplicates at import

Imports are where duplicates arrive in bulk. Before importing:

1. Dedupe the file itself by sorting on email and phone.
2. Normalize formats (lowercase emails, standard phone format).
3. Choose a duplicate rule for rows that match existing records:

Rule	Use it when
Skip	The CRM already has the best version, and the file is lower quality
Update	The file has newer or more complete information
Create	You’re sure the rows are different people despite a match (rare)

4. Test with a small batch and check the results before importing the rest.

Our CRM data migration guide covers imports end to end.

How to merge safely

Merging combines two records into one surviving record. Done carelessly, it can erase information. Set simple rules:

• Pick the survivor. Usually the older record (it has more history) or the most complete one.
• Field by field, keep the best value. Prefer the most recent phone and email; keep any value that only exists on one record.
• Keep all history. Notes, messages, calls, tasks, deals, quotes and files from both records should end up on the survivor.
• Keep the strictest consent. If either record has an opt-out for a channel, the merged record must keep it. Never let a merge turn messaging back on for someone who said stop.
• Keep the original source. The first-touch lead source is valuable for reporting; don’t overwrite it with the second record’s source.
• Reassign carefully. If the records have different owners, decide who keeps the relationship and tell the other rep.

Merge one at a time until you trust the process. Bulk merges are fast, but mistakes are fast too.

Data quality habits for small teams

You don’t need a data team. A few habits go a long way:

• Weekly five-minute review of flagged duplicates by one named person.
• Required fields kept to a minimum, so reps don’t invent values to get past a form.
• Consistent picklists for source, status and lost reason instead of free text.
• Cleanup before and after big imports.
• A quick audit each quarter: records with no owner, no email and no phone, or no activity in a year.

For more on keeping records useful, see our guide to contact management.

How Exa helps

Exa checks for duplicates as you work. Live duplicate detection matches records by email, phone and name, and you can merge duplicates when they’re found. CSV imports let you choose how matching rows are handled: skip, update or create. Every lead, contact, company and deal has an activity timeline, so the history a merge brings together stays in one place.

Exa also stores consent per channel with its source and evidence, and opted-out numbers are blocked server-side (security and compliance). Audit logs record creates, updates and deletes with old and new values and the user who made the change, so you can trace what happened if a record changes unexpectedly.

Final thoughts

You’ll never get duplicates to zero, but you can make them rare and harmless. Normalize email and phone, match on them first, stop duplicates at entry and import, merge with clear rules that protect history and consent, and give one person five minutes a week to keep things tidy. Clean records make every report, automation and follow-up more reliable.