How duplicate data quietly wrecks your CRM

The hidden cost of a duplicate row

A duplicate row looks harmless in a spreadsheet — just one extra line. Inside a CRM it behaves very differently. Each record is a container for emails, calls, deal stages, and ownership, so two records for the same person split that history in two. A rep opens one, sees no recent activity, and calls a lead a colleague spoke to yesterday. Reporting compounds the problem: if 18% of your contacts are duplicates, every count, conversion rate, and forecast built on those contacts is wrong in the same direction.

Four ways duplicates do damage

• Inflated metrics. “10,000 contacts” might be 8,200 real people. Pipeline value, list size, and per-rep quotas all drift from reality.
• Split activity history. Notes, emails, and deals scatter across copies, so no single record tells the whole story of the relationship.
• Double-fired automations. Two records for one person can trigger a welcome sequence twice, double-count in attribution, or assign the same lead to two reps.
• Embarrassing outreach. The same prospect receives the same “just checking in” email from two people, which reads as disorganized at best.

Why duplicates accumulate

Duplicates rarely arrive all at once. They build up through ordinary operations: a lead fills out two forms, a list is imported twice, a sales rep adds a contact that marketing already had, or two systems sync without a shared key. Most are not exact copies. They differ in tiny, machine-significant ways — a trailing space, a capitalized domain, a phone number written as (555) 123-4567 in one row and +15551234567 in another. To a person these are obviously the same; to a CRM’s exact-match rules they are two different people.

Exact vs. near-duplicates

What your CRM’s own deduplication catches

Most CRMs do run a duplicate check on import, which is why duplicate data in a CRM is often assumed to be a solved problem. It is worth knowing exactly how narrow that check is. HubSpot deduplicates contacts on email: if a contact already exists with the same email address, the imported information updates that contact instead of creating a new one [1]. Useful — and limited to one field. If your file has no email column, or no unique identifier at all, HubSpot treats every row as a new contact [1]. Two records for the same person under a work address and a personal one pass straight through, and phone-number format drift is not part of the check.

The second limitation is that cleaning up afterwards is one-way. HubSpot’s documentation states that it is not possible to unmerge records: the timeline activities and associated records of both sides land on the survivor, and there is no undo [2]. A wrong merge inside a live CRM is permanent. A wrong delete in a spreadsheet is one Ctrl+Z. That asymmetry, more than any efficiency argument, is why the dedupe pass belongs in the file.

Normalizing first is what makes the CRM’s own email match actually work. Lowercase and trim the address column with the email cleaner so “Jane@Acme.com ” and “jane@acme.com” stop counting as two people, and rewrite every phone number into E.164 so the fallback match has something consistent to compare. If the destination is HubSpot specifically, the pre-HubSpot import checklist lists the field-level rules the importer enforces, and the walkthrough for a Salesforce data import covers the equivalent ceilings on the Data Import Wizard and Data Loader.

Two more edges of that check are worth knowing, because they decide what your file has to look like. Companies are deduplicated on a different field entirely — HubSpot matches the primary value of the company domain name property rather than an email [1] — so a contact list and an account list need different columns normalised before either import. And when a row matches more than one existing record, the import does not pick a winner: it errors [1]. A file with three rows for the same person can therefore fail the import outright rather than quietly duplicating, which is a better failure but still a failure you fix in the spreadsheet.

Duplicates are a compliance problem, not only a reporting one

If the records are personal data, accuracy stops being a preference. Article 5(1)(d) of the GDPR requires personal data to be “accurate and, where necessary, kept up to date”, and that “every reasonable step must be taken” to erase or rectify inaccurate data without delay [3]. Two copies of the same contact that disagree — different job title, different consent state, different phone number — mean at least one of them is inaccurate by definition.

The practical risk is a request answered against the wrong copy. Someone asks to be removed or corrected, a rep finds the record the email match linked, updates it, and closes the ticket — while the second record, the one under a personal address that the duplicate check never caught, keeps its old values and keeps feeding the next campaign. The obligation is not discharged, and nobody knows.

That is a reason to do the cleanup somewhere it does not create a second exposure of its own. Uploading a contact export to a web-based cleaner to make it compliant hands the same personal data to another processor; why the upload step is the risky part works through what that actually costs you. Sigmera runs the pass in the browser, so the list never leaves the device — and the comparison of cleaning routes by where the file is processed sets that against the server-based and desktop options.

Fix it before the import, not after

The cheapest place to remove a duplicate is in the file, before it ever reaches the CRM. A row you delete from a spreadsheet leaves no trace; a record you merge inside a live CRM means reconciling ownership, activity, and automations that have already fired. Run a dedupe pass on the export, decide whether to keep the first or last copy of each match, normalize email case and phone format so near-duplicates collapse together, then import the clean file. The full pre-import checklist puts those steps in order; if the list is already open in a workbook, Excel’s built-in Remove Duplicates command handles the exact matches before you export. For a lead list specifically — matching on email, ignoring case and spacing, and choosing which copy survives — the sales-lead dedupe walkthrough covers the whole pass end to end.