Skip to content
English
Contact us
  • There are no suggestions because the search field is empty.

Match Incoming Records to Existing HubSpot Records

Confirm which existing companies and contacts each row belongs to

Overview

Matching connects incoming rows with companies and contacts that already exist in HubSpot. When it works well, imports enrich existing records instead of creating duplicates, and notes and associations land on the right records. The Studio matches conservatively. It accepts strong evidence automatically and shows uncertain cases for review.

What this helps you understand

  • Which incoming rows correspond to existing HubSpot companies and contacts.
  • Why a candidate was or was not accepted.
  • Which rows need a human decision before the import.

Before matching starts

Matching runs only when Look up existing HubSpot records before exporting is turned on in Details. The lookups are read-only.

Rows that already contain a valid HubSpot record ID use that ID directly and aren't searched again. A record ID must be a HubSpot number. Temporary IDs from other systems are rejected.

How contacts are matched

Contacts are matched by exact email address only.

  • Both the primary email and any additional emails on HubSpot contacts are checked.
  • Matching ignores letter case and surrounding spaces.
  • Names, fuzzy similarity and AI are never used to match contacts.
  • When a match is found, the prepared file uses the existing contact's primary email, so HubSpot updates the right record.
  • When several contacts share the email, the best match is selected and the row is flagged with a warning to review.

Rows without a valid email are not searched. They can still be created when they have a first and last name.

How companies are matched

Companies are often recorded without a domain, so the Studio uses wider evidence:

  • Similarity between the incoming and existing company names, taking word order and distinctive words into account.
  • Country, when available.
  • The company properties selected in Settings, such as a unique company ID.

Each candidate receives a match score. Only a strong candidate is selected automatically. When the selected properties contradict each other, such as different values for a unique identifier, the candidate isn't accepted automatically.

Rows need a company name to be searched. Repeated company names in the same file reuse the same search results.

Optional AI company matching

Turn on Enable AI company matching for this job in Details for company names that are hard to compare, such as abbreviations, trading names and local spellings.

  • The rules-based score still applies. AI adds extra evidence for difficult decisions.
  • Only the company properties selected in Settings and the relevant matching evidence are shared. Other columns and contact data are not.
  • If the AI service fails, the problem is shown on the row. It is never treated as "no match".

Confirm that AI use meets your organisation's data policies before turning it on.

Review matching decisions

In Review, open a row to see its Source-row decisions:

  • Company candidate: Choose a specific company from the candidates, each shown with its score, or keep the automatic selection.
  • Contact candidate: Choose a specific contact, shown by primary email and record ID, or keep the automatic selection.
  • Company record ID: Enter a known company ID directly.
  • Contact email override and Company domain override: Correct the identifier for this row.

Expand Source data & matching evidence to see the source values and candidate details behind each decision.

Decisions are saved with the import. Choose Apply changes & rerun to refresh the files.

What matching does not do

  • It does not merge existing duplicate records in HubSpot. Use Matching and Deduplication in app.furball.tech to find and resolve them.
  • It does not change existing records during preparation. Updates happen only when you import the files or confirm automated ingestion.
  • A lookup failure is shown as an error, never as a confirmed "no match".

Why it matters

Reliable matching keeps one record per real customer, so activity, ownership and reporting stay complete. Conservative rules prevent incoming data from attaching to the wrong account, which is harder to undo than a duplicate. Clear evidence lets your team make the final call on borderline cases.

What to do next

  • Review flagged rows in Review Results and Correct Source Rows.
  • Select useful unique company properties in Settings to strengthen company matching.