Zendesk to Salesforce Migration: 6 Steps Without Data Loss
Zendesk to Salesforce migration: map tickets to Cases, choose CaseComment, EmailMessage or FeedItem, size Bulk API loads, plan cutover. Read the steps.
Published Updated 12 min read

On this page
- Why Zendesk to Service Cloud projects stall on ticket history
- What moves, what gets rebuilt and what's at risk?
- How do Zendesk objects map to Salesforce?
- Which tool should load Zendesk data into Salesforce?
- Getting data out of Zendesk
- Throughput and timeline math
- Zendesk to Salesforce migration in 6 steps
- For Salesforce partners and AEs: getting the deal unstuck
- How MigrateX does this
- Zendesk to Salesforce migration checklist
- Zendesk to Salesforce migration FAQ
- Changelog
A Zendesk to Salesforce migration moves tickets into Service Cloud as Cases, with end users as Contacts, organizations as Accounts and agents as Users. The hard part is ticket history: each Zendesk comment has to land on the right Salesforce object (EmailMessage, FeedItem or CaseComment) with its visibility, author and date intact. Triggers, macros and SLA policies do not move; you rebuild them.
This guide is for support ops leads, Salesforce admins and Salesforce partners. It maps every object, compares the three import methods, sets Zendesk export limits beside Salesforce load limits, and ends with a cutover plan and checklist.
Why Zendesk to Service Cloud projects stall on ticket history
Accounts and Contacts are flat records that Salesforce's own tools import. Projects stall on ticket history, for three reasons. There is no one-to-one target for a Zendesk comment: three Salesforce objects can hold it, each with different limits. Original dates and authors don't carry over unless you enable audit fields. And the volume sits in the comments: a ticket with 30 replies is 31 records to write. Plan for history first.
What moves, what gets rebuilt and what's at risk?
Data moves. Configuration is rebuilt. Running SLA timers never move.
| Moves (data) | Rebuilt in Salesforce (configuration) | At risk (needs a plan) |
|---|---|---|
| Tickets, as Cases | Triggers and automations, as Flows and assignment rules | Original created dates and comment authors |
| Public comments and internal notes | Macros, as Salesforce Macros or Quick Text | Comments over 4,000 bytes (if you use CaseComment) |
| End users and organizations, as Contacts and Accounts | SLA policies, as entitlement processes and milestones | Closed dates, since Case ClosedDate is not createable |
| Agents, as Users | Views, as list views | Archived tickets missed by list endpoints |
| Groups, as Queues | Ticket forms and field conditions, as record types and page layouts | Running SLA timers, CSAT, side conversations |
| Custom field values and tags | Business hours | Ticket audit trail (field-by-field change history) |
| Attachments, as Files | Email channels, as Email-to-Case | Help Center inline images and article links |
Salesforce documents an entitlement process as "a timeline that includes all the steps (MilestoneType records) that your support team must complete to resolve cases", and its SlaProcess object supports query and describe calls only, not create (Salesforce Object Reference, as of 4 Nov 2026). So SLA policies are rebuilt in Setup, not loaded.
How do Zendesk objects map to Salesforce?
Object and field behaviour in this section is from the Salesforce Object Reference (Winter '27, API version 68.0), checked on 4 Nov 2026. The mapping choices are ours.
| Zendesk | Salesforce | Notes |
|---|---|---|
| Ticket | Case | Store the Zendesk ticket ID in an external ID field so loads can upsert |
| Ticket status | Case Status picklist | IsClosed "is controlled by the Status field; it can't be set directly" |
| Ticket priority | Case Priority | Map urgent, high, normal, low to your picklist values |
| Ticket channel (via) | Case Origin | Origin is "the source of the case, such as Email, Phone, or Web" |
| Requester (end user) | Contact (Case ContactId) | Load Contacts before Cases |
| Organization | Account (Case AccountId) | A user's default organization only; see the people section |
| Agent | User | Agents who left can be created as inactive Users |
| Group | Queue (a Group whose Type is Queue) | Link the queue to Case through QueueSobject |
| Assignee or group on a ticket | Case OwnerId (a User or a Queue) | OwnerId is polymorphic |
| Public comment | EmailMessage (ParentId = Case) | Recommended; see the comment section |
| Internal note | FeedItem, Visibility = InternalUsers | Recommended; CaseComment with IsPublished = false is the fallback |
| Attachment | ContentVersion, linked to the Case by ContentDocumentLink | Or set FirstPublishLocationId to the Case |
| Custom ticket fields | Custom Case fields | Recreate types; dropdowns become picklists |
| Tags | A text or multi-select picklist field | Salesforce Case has no standard tag field |
| Help Center article | Knowledge article (Knowledge__kav) | Needs Lightning Knowledge enabled |
| SLA policy | Entitlement + entitlement process + milestones | Rebuilt, not loaded |
Where should Zendesk comments go: CaseComment, EmailMessage or FeedItem?
This decision shapes the project. The trade-offs, from Salesforce's object reference (as of 4 Nov 2026):
| CaseComment | EmailMessage | FeedItem | |
|---|---|---|---|
| Body | Plain text, "maximum size of the comment body is 4,000 bytes" | HtmlBody and TextBody | Rich text, limited HTML tags |
| Visibility | IsPublished true or false | IsExternallyVisible | Visibility: AllUsers or InternalUsers |
| Original date | "contact Salesforce" to set audit fields such as CreatedDate | MessageDate is createable | CreatedDate settable "during migration" |
| Original author | Same as above | FromAddress and FromName; CreatedById must be the running user or blank on single inserts | CreatedById settable during migration |
| Editable after load | Only IsPublished | Only IsExternallyVisible and custom fields, once not a draft | Body and Visibility |
| Prerequisite | None noted | Email-to-Case or Enhanced Email | None noted |
Our recommendation:
- Public comments → EmailMessage. It keeps the HTML body, the direction (Incoming true for customer messages) and the original MessageDate; set Status to 3 (Sent) for agent replies. A sent EmailMessage can't be edited later apart from IsExternallyVisible and custom fields (EmailMessage reference), so get the transform right in the test migration.
- Internal notes → FeedItem with Visibility InternalUsers. The FeedItem reference says that during migration "the context user can set these fields to the original post's author and creation date. The fields can't be updated after migration" (FeedItem reference). Rich text allows only p, a, b, code, i, u, s, ul, ol, li and img tags (no br), so convert line breaks and flatten tables.
- CaseComment only when you must. IsPublished maps cleanly to Zendesk's public flag, but the 4,000-byte cap splits long notes and setting CreatedDate needs a request to Salesforce (CaseComment reference). Choose it if your reports or portal already use case comments.
Dates, authors and owners
Original dates need the "Set Audit Fields upon Record Creation" permission, which lets users set CreatedDate, CreatedById, LastModifiedDate and LastModifiedById "only upon record creation — future edits to those fields are not permitted" (Salesforce Help article 000386875, as of 4 Nov 2026). Enable it in User Interface settings with "Update Records with Inactive Owners", needed when tickets belong to agents who have left. Case ClosedDate lists no Create property, so keep Zendesk's solved and closed timestamps in custom date fields.
People, files and knowledge
For users in several Zendesk organizations, pick a rule (default organization, or Contact plus Account Contact Relationships) before loading Contacts.
Attachments become Files. Setting FirstPublishLocationId to the Case creates the file and shares it with the record "in a single transaction" (ContentVersion reference). Set ContentDocumentLink Visibility to InternalUsers for files on internal notes.
Knowledge is a licence question: Salesforce's Lightning Knowledge Guide (Winter '27) lists it in Essentials and Unlimited Editions with Service Cloud, and "for an additional cost" in Professional, Enterprise, Performance and Developer Editions. Its Import Articles tool takes a .zip of up to 20 MB with a .csv of at most 10,000 rows, header included, and 32 KB per cell.
Which tool should load Zendesk data into Salesforce?
Salesforce documents two native import tools; the third route is the API.
| Method | Record limit | Imports Cases? | Best for |
|---|---|---|---|
| Data Import Wizard | Up to 50,000 | No | Accounts and Contacts in small orgs |
| Data Loader | Between 5,000 and 5 million | Yes | One-off loads by an admin |
| Bulk API 2.0 and REST (scripted or managed) | 150,000,000 records per rolling 24 hours | Yes | Full history with files, comments and delta sync |
Sources: Salesforce Help, choosing a method for importing data and Bulk API limits, as of 4 Nov 2026.
Method 1: Data Import Wizard
The wizard is in-browser and imports "accounts, contacts, leads, solutions, campaign members, and custom objects", up to 50,000 records at a time. Cases are not on that list, so it can't carry tickets.
Use it for: Accounts first, then Contacts matched to Accounts, from Zendesk CSVs. De-duplicate end users with several email identities on their primary email first.
Method 2: Data Loader
Data Loader is a desktop application that can "insert, update, delete, or export Salesforce records", for between 5,000 and 5 million records. Salesforce lists it for Enterprise, Unlimited, Performance, Developer and Database.com Editions.
Steps: create external ID fields; load Accounts, Contacts and Users; upsert Cases on the Zendesk ticket ID; then load comment rows that reference each Case. Limits: you flatten Zendesk JSON into CSVs, split long comments and handle files separately. Delta sync is a manual re-run. Troubleshooting: failed rows land in an error file. Check load order, missing picklist values and auto-response rules: with one active, "SuppliedEmail is required when creating a case via the API" (Case reference).
Method 3: Bulk API 2.0 and REST, scripted or managed
For full history, load records through Bulk API 2.0 and files through REST, scripted or by a migration service.
POST /services/data/v68.0/jobs/ingest
Content-Type: application/json
{
"object": "Case",
"operation": "upsert",
"externalIdFieldName": "Zendesk_Ticket_ID__c",
"contentType": "CSV",
"lineEnding": "LF"
}
PUT /services/data/v68.0/jobs/ingest/{jobId}/batches (CSV body)
PATCH /services/data/v68.0/jobs/ingest/{jobId} {"state": "UploadComplete"}
GET /services/data/v68.0/jobs/ingest/{jobId}/failedResults
Bulk API 2.0 accepts 150 MB of base64-encoded data per job; Salesforce advises uploading no more than 100 MB because base64 adds about 50%. Bulk API and Bulk API 2.0 share an allocation of 15,000 batches per rolling 24 hours (Bulk API limits, as of 4 Nov 2026).
Files go through the REST sObject resource as multipart requests:
POST /services/data/v68.0/sobjects/ContentVersion
Content-Type: multipart/form-data; boundary=boundary
--boundary
Content-Disposition: form-data; name="entity_content"
Content-Type: application/json
{"Title": "invoice.pdf", "PathOnClient": "invoice.pdf", "FirstPublishLocationId": "500..."}
--boundary
Content-Disposition: form-data; name="VersionData"; filename="invoice.pdf"
Content-Type: application/pdf
(binary)
--boundary--
Multipart uploads take files up to 2 GB for ContentVersion; a non-multipart request is limited to 50 MB of text or 37.5 MB of base64-encoded data (REST API blob guide, as of 4 Nov 2026).
Troubleshooting: you own retries and ordering. Keep a Zendesk ID to Salesforce ID ledger so failed batches replay without duplicates. The QueueSobject reference warns: "You can't update or insert more than 18 queues at once when using the Bulk API."
Getting data out of Zendesk
The export side sets the pace. Zendesk's incremental export endpoints allow 10 requests per minute, or 30 with the High Volume API add-on, and return up to 1,000 records per page (Zendesk incremental exports and rate limits, as of 4 Nov 2026).
GET /api/v2/incremental/tickets/cursor?start_time=1609459200
GET /api/v2/incremental/ticket_events?start_time=1609459200&include=comment_events
GET /api/v2/tickets/{ticket_id}/comments (max 100 per page)
GET /api/v2/help_center/incremental/articles?start_time=1609459200
Without the comment_events sideload, ticket events flag comments only as comment_present and comment_public. The per-ticket comments endpoint uses the general Support API limit: 200 requests per minute on Team up to 2,500 on Enterprise Plus or with High Volume.
Two traps. Zendesk archives tickets 120 days after they close, and the Listing Tickets, Listing Ticket Metrics and Listing Ticket Audits endpoints don't return archived tickets (Zendesk archiving article, as of 4 Nov 2026), so build the inventory from the incremental export. And authentication is changing: Zendesk API tokens "will be permanently deactivated on April 30, 2027", and since 27 Oct 2026 existing accounts can no longer create new tokens (Zendesk OAuth migration guide, as of 4 Nov 2026). Build the extractor on OAuth.
Our guide on how to export tickets from Zendesk covers the export routes in more depth.
Throughput and timeline math
Comments and files set the duration, not tickets. Work out each side's ceiling; the slower one wins.
Example: a Zendesk account with 200,000 tickets. The incremental tickets export at 1,000 per page and 10 requests per minute lists them in about 20 minutes. Pulling comments per ticket at a 700-per-minute Enterprise limit takes at least 200,000 requests, about 5 hours, before paging or retries. On the Salesforce side, a daily cap to check early is Files: Salesforce allows 200,000 ContentVersion publishes per 24 hours (ContentVersion reference), so 600,000 attachments need at least three days of loading.
Add time for at least one full test migration; our test migration checklist lists what to sign off. Migrated Cases, emails and files also use Salesforce data and file storage.
Zendesk to Salesforce migration in 6 steps
- Inventory and scope. Count tickets, comments, attachments and articles by year, archived included. Our guide to migrating ticket history sets out how much to move.
- Map and decide. Fix the object mapping, comment target, status and priority rules, and the rebuild list.
- Prepare Salesforce. Create Users, Queues, record types and external ID fields, enable audit fields, and switch off rules, Flows and notifications for the load.
- Test migration. Load the longest tickets, largest files and most internal notes. Check what a customer sees in your portal.
- Full load, then delta sync. Load history while Zendesk stays live, then upsert tickets changed since the last cursor. Freeze Zendesk only for the final delta.
- Validate and go live. Reconcile counts by status, owner and year, then sample comment order, authors, dates, visibility and files; our data migration validation guide covers sign-off. Keep Zendesk read-only until then.
For Salesforce partners and AEs: getting the deal unstuck
When a Service Cloud deal depends on moving Zendesk history, the buyer's blocking question is rarely "can Salesforce do this?" It is "who carries the risk of losing our ticket history?"
Three things unstick it. Answer the comment-mapping question in writing, using the table above. Offer a test migration on the buyer's own tickets in a sandbox before signature. And name an owner for cutover and validation: the SI, a migration specialist or the customer's admins. An unowned migration stalls the deal, not the data volume.
How MigrateX does this
MigrateX runs extraction, mapping, test migration, delta sync and validation as one managed project. We are SOC 2 Type II, GDPR compliant and HIPAA compliant, with a DPA and BAA available (security).
Salesforce Service Cloud is a supported destination among our help desk platforms. Moving the other way? The Salesforce Service Cloud to Zendesk route lists what moves. For a sense of pace, one MigrateX project moved 70,000 tickets in 24 hours.
Start with a free 100-record sample of your own Zendesk data and get a fixed quote within 24 hours: contact us.
Zendesk to Salesforce migration checklist
- Confirm licences for every agent who needs a User, and Knowledge licensing if Guide is in scope.
- Switch the Zendesk extractor to OAuth.
- Decide the comment target per comment type and visibility.
- Create external ID fields, plus custom fields for original solved and closed dates.
- Enable "Set Audit Fields upon Record Creation" and "Update Records with Inactive Owners".
- Create Queues (no more than 18 per Bulk API call) and link them to Case.
- Turn off assignment rules, auto-response rules, Flows and notifications.
- Rebuild triggers, macros, views and SLA policies.
Facts checked November 4, 2026 from Salesforce Help: choosing a data import method, Salesforce Object Reference: CaseComment, Salesforce Bulk API 2.0 limits, Zendesk incremental exports API and Zendesk OAuth migration guide. No vendor paid for this article. How we research
Zendesk to Salesforce migration FAQ
Can I use the Data Import Wizard to move Zendesk tickets to Salesforce?
No. Salesforce lists accounts, contacts, leads, solutions, campaign members and custom objects, up to 50,000 records, and not Cases. Use it for Accounts and Contacts, and Data Loader or the API for Cases and comments.
Do Zendesk comments become case comments in Salesforce?
They can, but CaseComment caps each body at 4,000 bytes and needs Salesforce's help to set original dates. Many teams put public replies in EmailMessage and internal notes in FeedItem with Visibility set to InternalUsers, which keep HTML, direction and original dates.
Will original ticket dates and authors be kept?
Only if you plan for it. Enable "Set Audit Fields upon Record Creation" to set CreatedDate and CreatedById at insert; they can't be changed later. FeedItem supports this during migration, EmailMessage carries the date in MessageDate, and Case ClosedDate is not createable, so store it in a custom field.
Do Zendesk triggers, macros and SLA policies migrate?
No. They are configuration. Triggers and automations become Flows and assignment rules, macros become Salesforce Macros or Quick Text, and SLA policies become entitlement processes with milestones, which Salesforce's API lets you query but not create.
How long does a Zendesk to Salesforce migration take?
There is no neutral public benchmark. It depends on comment and attachment volume, Zendesk's 10-per-minute incremental export limit, Salesforce's 200,000 file publishes per 24 hours, and the number of test runs.
Changelog
- 15 Mar 2026: First published as "Why Data Migration Stalls Help Desk Deals".
- 4 Nov 2026: Rewritten and moved from /blog/why-migration-kills-deals. Now a Zendesk to Salesforce Service Cloud guide: object mapping, comment targets, import methods, limits checked against vendor docs, and a cutover plan. The deal-stall material is one short section.
Written by
Ashwin VasudevanFounder & CEO, MigrateX
Ashwin built MigrateX after watching enterprise deals stall over migration fear. He's spent the last decade helping platform vendors and channel partners close deals by eliminating data migration risk.


