Why does a Pipedrive to Attio migration create duplicate companies?

Because Pipedrive organizations are identified by name, not by domain. Attio matches companies on domain, so an export whose organization rows have no domain column gives the importer nothing to match on and every organization is created — including the ones that already exist in Attio, and the ones Pipedrive itself held twice under slightly different names.

How do Pipedrive objects map to Attio?

Organizations become Attio companies, persons become Attio people, and deals become Attio deals. Load them in that order — companies, then people, then deals — so associations resolve against records that already exist rather than creating new ones as a side effect.

What should be done about Pipedrive persons with multiple email addresses?

Decide which address is primary before the import, because Attio matches people on the email attribute. If the export puts a personal address first and Attio already holds the work address, the row will not match and a second person record is created. Keeping both addresses on the record is fine; the question is which one the import matches on.

Should duplicates be cleaned in Pipedrive or in Attio?

In Attio, in most cases. Pipedrive's duplicates arrive with you either way, and cleaning after the import means you only learn one tool's rules and can match on the identifiers Attio actually uses. The exception is if the migration is weeks away and someone is already doing the work in Pipedrive — a cleaner source is always a cleaner import.

Duplicates after a Pipedrive to Attio migration

Last checked September 1, 2026 · Deduptio is our product. Every claim about another tool comes from its public site or documentation and is linked in place — tell us if something is out of date.

The short answer

Pipedrive identifies organizations by name; Attio matches companies by domain. Add a domain column to the organization export before importing, pick one primary email per person, load organizations before people before deals, and turn off mailbox auto-creation until it is finished.

The structural mismatch

Most migration problems are not bugs, they are two data models disagreeing. Here the disagreement is specific and it explains nearly every duplicate you will see:

Pipedrive	Attio	Consequence
Organizations identified by name; domain is optional and often empty	Companies matched on domain	Nothing to match on, so every organization row creates a record — including ones that already exist
Persons can hold several email addresses, one flagged primary	Matching happens on the email attribute you map	Map the wrong address and the row matches nothing, creating a second person record
Free-text organization names, entered by reps over years	Fuzzy matching available, but not automatic	"Acme", "Acme Ltd" and "ACME Limited" arrive as three companies
Deals linked to an organization and a person	Deals reference company and person records	Duplicate parents produce duplicate-looking deals; clean parents first

Before you import

Add a domain column to the organizations export. If Pipedrive holds a website field, derive the domain from it and normalize: lowercase, no protocol, no www., no trailing slash. This single column is the difference between an import that matches and one that duplicates.
Collapse obvious organization-name variants in the file while you are in a spreadsheet — this is the one stage where a spreadsheet is exactly the right tool, because the records have no history to protect yet.
Choose one email per person as the matching attribute, preferring the work address. Keep the others as additional values if your mapping allows.
Turn off mailbox auto-creation until the import is done, or Attio will create people from inbound mail alongside the ones you are loading.
Load in order: organizations → persons → deals, mapping the unique attribute at each step, and test with twenty rows first.

After the import: the cleanup order

Companies, exact. Same normalized domain. Safe to merge in bulk after a sample check.
Companies, reviewed. Fuzzy name plus one exact signal — same country, same LinkedIn URL, or associated people sharing an email domain. This is where the free-text organization names get resolved; expect to reject some, and read fuzzy matching company names before you start.
People, exact. Same email. Catches the mailbox-sync collisions and any double-run rows.
People, reviewed. Name plus company, or name plus phone. Now meaningful, because the companies are unique.
Deals. Last, once their parents are clean — many apparent duplicate deals simply disappear when two company records become one. See duplicate deals in Attio.
Lists. Merging records that shared a list leaves the survivor in it twice.

The Pipedrive-specific things to check afterwards

Organizations that were people. Sole traders often exist as both an organization and a person in Pipedrive. In Attio that is a company and a person, which is correct — do not merge across objects trying to tidy it.
Deal owners. Merging a company can leave deals pointing at the right record with the wrong owner if two reps each kept their own version of the account.
Lost-then-recreated deals. Pipedrive pipelines accumulate these; they look like duplicates and are often legitimate history. Close one as a duplicate rather than merging if you are unsure.

Scan the new workspace before the team moves in

Connect your Attio workspace, write one match rule, and run a scan. Scans are read-only — you see every group and the reason it matched before anything is merged.

Start a free Attio duplicate scan