Migration guides

HubSpot to Zendesk Migration: Move Support, Keep the CRM

HubSpot to Zendesk migration for support only: what leaves HubSpot by which route, what Zendesk's import can set, and how contacts stay matched.

Ayush BothraFounding Member, MigrateX

Published 9 min read

Clay blocks for HubSpot and Zendesk: ticket cards travel to Zendesk while a stack of deal cards stays beside HubSpot, for a HubSpot to Zendesk migration
On this page

A HubSpot to Zendesk migration can move support only and leave sales in HubSpot. Tickets, the emails and notes on them, inbox threads and knowledge base articles move to Zendesk through its Ticket Import and Help Center APIs. Contacts and companies stay in HubSpot, and Zendesk receives a copy keyed by external_id.

This guide is for support and operations leaders who have chosen Zendesk for the service team while sales and marketing stay on HubSpot. Still deciding whether to split at all? Start with CRM vs help desk.

What moves in a HubSpot to Zendesk migration, and what stays?

Support records move, customer records are copied, and sales data stays. No MigrateX route page covers this direction, so the full mapping is here, as of November 2026. "Fixed" means the two platforms decide it. "Our mapping" is our design choice, and "Your choice" is yours; both are agreed on your free sample.

HubSpot recordLands in Zendesk asWhat the import keepsFixed or your choice
TicketTicket, through the Ticket Import APICreated, updated and solved dates; requester; assignee; HubSpot record ID in external_idFixed
Pipeline stageTicket statusHubSpot marks each stage OPEN or CLOSED, the starting point for open or solvedYour choice: each stage mapped to a status
Email activity on a ticketPublic commentAuthor, original time, bodyOur mapping, agreed on your sample
Note on a ticketInternal noteAuthor, original time, bodyFixed
Call, meeting or taskInternal note, or left behindTime and ownerYour choice
Inbox or help desk threadComments on the ticket the thread belongs toEach message's sender and time; thread comments as internal notesOur mapping for threads on a ticket; your choice for threads without one
AttachmentAttachment on the matching commentFile, uploaded before the commentFixed
Knowledge base articleHelp Center articleBody and author; publish date does not carry overFixed; category and section layout your choice
ContactEnd userEmail, name, HubSpot record ID in external_idYour choice: all contacts, or those with a ticket
CompanyOrganizationName, domains, HubSpot record ID in external_idYour choice, as for contacts
Survey responseTicket field or tagScore and comment as valuesFixed that it is not a rating; field or tag your choice
Deal, marketing dataNot loadedStays in HubSpotFixed
Workflow, SLA, view, reportRebuilt by your Zendesk adminNothingFixed

Two rows are fixed by vendor rules. HubSpot's Conversations API says thread comments "only appear in the inbox and are not sent to visitors", so they arrive as internal notes. Zendesk's Satisfaction Ratings API says only the ticket's requester can create a rating, so HubSpot survey scores cannot become Zendesk ratings.

How does ticket history leave HubSpot, and through which route?

Through the APIs. HubSpot's built-in exports each leave out part of the history, so they serve as counts and safety copies, not as the source of the move.

HubSpot routeWhat it holdsLimit, as of November 2026Role in this move
Record export (UI)Ticket properties as rowsNo activities such as calls and notes; 300 exports a rolling 24 hoursCounts to reconcile against
Exports APITickets with up to four associated objects, as record IDs30 exports a rolling 24 hours, one at a timeTicket-to-contact and ticket-to-company IDs
CRM backupTickets, contacts, companies, conversationsNo associations or activity dataSafety copy
CRM object and associations APIsTickets, contacts, companies, emails, notes and their links100 records per batch; 1,000 inputs per association read; 625,000 calls a day on ProfessionalThe main read
Conversations APIInbox and help desk threads, message by message500 threads per page; email replies truncatedThread history
Knowledge base exportArticle data such as title, language and URLService Hub Professional or Enterprise, with a Service Hub seatThe article list to check against

Two details decide how long the read takes:

  • Email messages arrive shortened. HubSpot says it "cuts off the reply history section of an email". Each message carries a truncationStatus, and any message marked truncated is read again from the original content endpoint.
  • Threads are found through their ticket. The Conversations API filters threads by associatedTicketId, so each ticket's threads are read with it, and threads with no ticket are listed separately.

The daily allowance is shared with your other HubSpot integrations. Every export route and what it leaves out is compared in 9 ways to export HubSpot data.

What can a Zendesk import set on each record?

Original dates, authors and visibility, but not SLA history or article dates. Zendesk's Ticket Import API, open to admins only, takes one ticket or up to 100 per bulk call.

Zendesk recordThe import can setLimits, and what Zendesk sets itself
Ticketcreated_at, updated_at, solved_at, requester, assignee, tags, external_idZendesk assigns the ticket id. No metrics or SLAs on imported tickets
Commentauthor_id, created_at, public, attachment tokensNo dates before 1970 or in the future. At most 5,000 comments per ticket
End userEmail, name, external_id; skip_verify_email stops verification emailsBulk user imports are off until Zendesk Customer Support enables them
OrganizationName, domain_names, external_id; 100 per bulk callNo matching on name when updating
Help Center articleHTML body, author_id; notify_subscribers: false stops emailscreated_at is read-only. Attachments up to 20 MB each
Satisfaction ratingNothingOnly the requester can create one

Imports are quiet by default. Zendesk says triggers won't run on an imported ticket that is not closed, until the ticket is next updated, and users copied on imported tickets get no notifications. A primary email is written only when the user is created: on a later update Zendesk adds a secondary address instead, so the email each user is created with stays primary.

How do contacts and companies stay matched after the split?

With one key on each side, written during the move. Both platforms document a field for it:

HubSpotZendesk
PeopleEmail, or a unique custom property matched through idPropertyEmail or external_id; the API treats external_id as case-insensitive
CompaniesCompany domain name, or a unique custom propertyid or external_id; Zendesk's Create Or Update Organization does not match on name

We map it as follows, agreed on your free sample: the HubSpot record ID goes into external_id on each Zendesk user and organization, and the Zendesk ID goes into a unique HubSpot property. Either side can then find the other without relying on names.

Two vendor-published integrations connect the platforms afterwards:

  • Zendesk, by HubSpot. HubSpot's App Marketplace lists an app named Zendesk, described as "Data Sync - Sync Zendesk with HubSpot data" and built by HubSpot. It syncs Zendesk users with HubSpot contacts and organizations with companies in both directions; ticket sync is one-way, and archived tickets do not sync. The listing requires Zendesk Suite Team, Growth or Professional, and custom field mappings need Data Hub Starter or above. HubSpot's data sync guide says it matches contacts by email and companies by name or domain by default.
  • Zendesk Support for HubSpot, by Zendesk. Zendesk's help centre says this integration shows ticket events and chat transcripts on the HubSpot contact timeline, flows one way from Zendesk to HubSpot, and syncs only the user's email address. It creates three triggers and one webhook in Zendesk.

Neither moves the history. We load history first and connect an integration afterwards, so the integration matches the emails and domains the migration loaded. The checks to run before connecting any CRM and help desk are in 7 checks before you connect a CRM and a help desk.

Facts checked November 17, 2026 from HubSpot: Export your records, HubSpot: Export your content and data, HubSpot: Back up CRM data, HubSpot: Export web content and data, HubSpot developer docs: Conversations API, HubSpot developer docs: API usage guidelines and limits, HubSpot developer docs: CRM Search API, HubSpot developer docs: Using object APIs, Zendesk developer docs: Ticket Import, Zendesk developer docs: Tickets, Zendesk developer docs: Ticket comments, Zendesk developer docs: Attachments, Zendesk developer docs: Users, Zendesk developer docs: Help Center articles, Zendesk developer docs: Satisfaction ratings, Zendesk developer docs: Rate limits, HubSpot developer docs: Exports, HubSpot developer docs: Associate records, HubSpot developer docs: Pipelines, HubSpot developer docs: Notes, HubSpot developer docs: Survey responses, HubSpot: Format import files, HubSpot: Connect and use HubSpot data sync, HubSpot App Marketplace: Zendesk (Data Sync by HubSpot), Zendesk help: Setting up the Zendesk Support for HubSpot integration, Zendesk developer docs: Organizations and Zendesk developer docs: Article attachments. No vendor paid for this article. How we research

HubSpot to Zendesk migration FAQ

Does anything change in HubSpot for the sales team?

No. Contacts, companies, deals and marketing data are read, not moved or deleted, and sales keeps working as before. The only write to HubSpot is optional: a property holding each record's Zendesk ID, if you want the two systems keyed to each other. Closing the Service Hub ticket pipelines afterwards is your decision.

Can HubSpot's Zendesk data sync app do the migration instead?

Not the support history. HubSpot's listing names users, organizations and tickets, says ticket sync is one-way and that archived tickets will not sync. It does not list emails, notes, inbox threads or knowledge base articles, which are the history this move is about. It suits keeping contacts and companies aligned after the load.

Should every HubSpot contact be copied into Zendesk?

It is a choice. Copying only the contacts who have raised a ticket keeps Zendesk small and avoids thousands of marketing leads appearing as end users. Copying everyone means an agent can find any customer on day one. Decide after you see how many contacts have no ticket at all.

Will the export use up our HubSpot API limit?

It shares it. HubSpot's usage guidelines give a Professional account 625,000 API calls a day for privately distributed apps and an Enterprise account 1,000,000, counted per account, so the read runs alongside your existing integrations. The run is paced to leave room for them, and the pace is set out with your quote.

Will customers get emails when old HubSpot tickets load into Zendesk?

Not from the import itself. Zendesk says triggers do not run on imported tickets that are not closed, and users copied on imported tickets get no notifications. New end users can be created with skip_verify_email, and article creation can suppress subscriber emails. Notifications resume when an imported ticket is next updated.

Can the HubSpot knowledge base be exported as a file instead?

Partly. HubSpot's knowledge base export produces a CSV, XLS or XLSX file on Service Hub Professional or Enterprise, and HubSpot says it includes article data such as title, language and URL. That is a list of articles, not their content, so each article is read and recreated in Zendesk sections and categories.

Get your HubSpot split checked before you commit

MigrateX is a managed data migration service on its own AI migration platform, SOC 2 Type II, GDPR and HIPAA compliant. Send us your ticket, thread and article counts and which contacts Zendesk should hold. You get 100 of your own tickets in Zendesk, free, and a fixed quote within 24 hours. Start with a free sample. Scope for each side is on the HubSpot and Zendesk pages, and the reverse move is Zendesk to HubSpot.

Written by

Ayush Bothra

Founding Member, MigrateX

Ayush writes about ITSM platforms, pricing and what it takes to move a service desk to a new platform without losing its history.