The moment you cancel Zendesk, you lose access. You can't log in. You can't run exports. You can't call the API.
Ninety days after that, Zendesk permanently deletes everything: tickets, conversations, custom fields, users, organizations. Per their own Service Data Deletion Policy, the process is irreversible once it begins.
Most "switching from Zendesk" guides spend 2,000 words comparing Freshdesk and Intercom feature tables. They mention data export in a single bullet point at the bottom. This guide reverses those priorities.
Why teams are leaving Zendesk in 2026
Three reasons dominate every conversation:
Pricing math that compounds. The Suite Professional base rate runs around $115/agent/month. Add Zendesk's Advanced AI add-on at $50/agent/month, and the per-resolution charge of $1.50 per automated resolution on top, and a 15-agent team's annual spend lands above $40,000 before you touch workforce management or quality assurance add-ons. For teams using less than 30% of available features — which is most teams — this stops making sense.
Implementation complexity that never gets simpler. Zendesk's admin surface spans multiple consoles: Admin Center for configuration, Explore for analytics, a separate console for AI. 73% of teams report 4+ week implementation times. Those aren't onboarding weeks — that's ongoing cognitive overhead for every configuration change.
AI is priced as a separate product, not a feature. You pay the AI add-on per agent and a per-resolution charge when the AI actually works. If your AI Agents deflect 5,000 tickets per month, that's $7,500/month on top of your base subscription. The incentive structure is backwards.
None of that is news to you if you're reading this. What you need now is the checklist.
Before you cancel: the data you're about to lose
What Zendesk deletes
The 90-day deletion window covers every piece of structured service data in your account:
- All tickets, including closed and archived
- All ticket comments — public replies and internal notes
- All custom field values (ticket, user, and organization fields)
- All user and organization records
- All ticket events and audit history
- All Sunshine Custom Objects and relationship records
The 90 days sounds like a cushion. It isn't. Zendesk does not guarantee reactivation will be approved, and reactivation requires resuming billing. Treat the cancellation date as your hard deadline for extraction.
What Zendesk's native export misses
Zendesk gives you a point-and-click export under Admin Center → Account → Tools → Reports. It produces CSV or JSON files. The JSON export is more complete than the CSV, but both have gaps:
Coverage
What Zendesk's native export actually includes.
| CSV export | JSON export | API | |
|---|---|---|---|
| Ticket subject, status, priority | ✓ | ✓ | ✓ |
| Ticket comments / conversation thread | ✗ | ✗ | ✓ |
| Custom ticket field values | Partial | Partial | ✓ |
| Multi-line text fields | ✗ | ✗ | ✓ |
| Multi-select fields | ✗ | ✗ | ✓ |
| Custom user fields | ✗ | ✗ | ✓ |
| Custom organization fields | ✗ | ✗ | ✓ |
| Sunshine Custom Objects | ✗ | ✗ | ✓ (before July 1) |
| Ticket attachments | ✗ | ✗ | ✓ |
The native exports are adequate for a quick ticket count audit. They are not adequate if you want to actually migrate your data somewhere, query it later, or feed it into an LLM.
Note: the built-in export is only available on Suite Growth, Professional, and Enterprise plans. If you're on Suite Team, you need the API regardless.
The July 1, 2026 Sunshine Custom Objects deadline
This one is separate from your cancellation timeline and affects you regardless of when you cancel.
Zendesk is shutting down the Sunshine Custom Objects API on July 1, 2026. After that date, the API returns 410 Gone. Even if you remain a paying Zendesk customer, those records become inaccessible.
If your Zendesk instance uses Sunshine Custom Objects — common for B2B SaaS companies storing hardware assets, license entitlements, account hierarchies, or any relational data that doesn't fit the standard ticket schema — you must extract those records before July 1, 2026, regardless of your cancellation timeline.
The extraction is non-trivial. Sunshine Custom Objects have a graph structure: object types, object records, and bidirectional relationship records. A naive export that grabs object records but skips relationship records produces data that can't be reconstructed. See our Sunshine Custom Objects extraction guide for the full schema and API sequence.
How to extract your Zendesk data
You have three options. The right one depends on your ticket volume and technical appetite.
Option 1: Evicta (under 24 hours, any volume)
Evicta connects to your Zendesk account with a read-only API token, runs a 100-record preview so you can validate the field mapping before anything is committed, then extracts your full instance to Postgres or JSONL.
What's included:
- All tickets via the Incremental Export API (handles rate limits, cursor state, and the documented ghost-cursor edge case automatically)
- All ticket comments, preserving the public/internal distinction
- All custom field values with LLM-assisted human-readable column names —
custom_field_8321becomescustomer_tier— reviewed and approved by you before any write happens - All users and organizations
- All Sunshine Custom Objects and relationship records (Premium tier)
- All attachments as S3-compatible URLs (Premium tier)
Pricing: $499 Starter (up to 50,000 tickets, Postgres + CSV, no Sunshine), $1,499 Premium (unlimited, Postgres + JSONL, Sunshine included).
The 100-record preview is free. No card required until you approve the full extraction.
Option 2: Build the extraction yourself
When this makes sense: under 50,000 tickets, no Sunshine Custom Objects, fewer than 10 custom fields, and at least one engineer who has worked with the Zendesk API before.
Realistic timeline: 2–4 weeks of focused engineering. The tricky parts are the Incremental Export API's cursor state machine (which has documented but rarely-documented-well edge cases), custom field name resolution (field IDs are anonymous integers by default), and the rate limit backoff logic.
The Incremental Export API cursor edge cases we've documented in detail in our Zendesk API extraction guide. Read that before starting.
Option 3: Migration agency (Help Desk Migration Service, Relokia)
When this makes sense: you want a fully managed extraction and import, end to end, with someone else taking accountability for the outcome.
Pricing: $3,000–$15,000 for mid-size migrations. The price scales with record count, custom field complexity, and destination platform.
What they're good at: white-glove project management, multi-platform expertise, taking the call when something goes wrong. What they're not good at: speed (typical engagement is 4–8 weeks) and custom field automation (most mapping is still manual).
Comparison
Extraction options at a glance.
| DIY script | Migration agency | Evicta | |
|---|---|---|---|
| Cost | Engineer-weeks | $3k–$15k | $499–$1,499 flat |
| Time to extracted data | 2–4 weeks | 4–8 weeks | <24 hours |
| Ticket comments included | If you write the API client | Yes | Yes |
| Custom field names resolved | Manual | Manual | Automated, you approve |
| Sunshine Custom Objects | If you write the API client | Requires custom SOW | Yes (Premium) |
| Data persisted on vendor servers | n/a | Often | Zero bytes |
How to cancel Zendesk
Once your data is extracted and validated, cancellation is straightforward. You must be the account owner — billing admins cannot cancel.
Standard accounts (self-serve billing)
- Log in to Zendesk Admin Center
- Click Account in the left sidebar
- Go to Billing → Subscription
- Click Manage
- Scroll to the bottom and click Remove products or cancel account
- Follow the prompts
Cancellation takes effect at the end of your current billing cycle. You retain access until then.
Sales-assisted accounts (enterprise contracts)
Contact your Zendesk account representative directly. Cancellation terms are governed by your contract. Allow at least 30 days notice — most enterprise contracts require it.
Timing recommendation
Don't cancel the same day you complete your extraction. Run your new help desk in parallel for at least two weeks. Route new tickets to the new system. Let old tickets close in Zendesk. Validate that the extracted data looks correct in your destination before you pull the trigger on cancellation.
Two weeks of parallel running costs you approximately two weeks of the cheaper platform's base subscription. It is vastly cheaper than discovering a data gap after your Zendesk account goes dark.
Where to go next
This guide intentionally doesn't spend much time on destination platform selection — that's a longer decision that depends on your team size, channel mix, and pricing sensitivity. Our Zendesk Alternatives guide covers the main options with honest trade-offs.
The short version: Freshdesk is the most common destination (cheapest omnichannel feature parity, strong migration tooling). Intercom is the best choice if you're moving toward a messenger-first, AI-deflection model. Pylon and Plain are strong for B2B SaaS teams with Slack-heavy customer bases. Salesforce Service Cloud is the right move if you're consolidating into Salesforce CRM and have the engineering bandwidth for the import.
Our destination-specific extraction guides if you know where you're going:
- Zendesk to Salesforce migration
- Zendesk to Pylon or Plain migration
- Zendesk to Intercom, Front, and Help Scout: the migration trap to avoid
When you should not use Evicta
Three honest scenarios:
You have under 5,000 tickets, no Sunshine Custom Objects, and no custom fields. Zendesk's native JSON export is sufficient. Download it, validate the ticket count, done. Don't pay anyone.
You want someone to manage the full migration end to end, including destination platform setup, data import, user training, and go-live cutover. Hire a migration agency. They exist for exactly this. Evicta does the extraction side; agencies do the whole project.
Your security policy requires that no data leave your own infrastructure, even transiently. The Premium tier includes a Zero-Trust CLI bundle (evicta-zero-trust-cli). Evicta generates the extraction logic and delivers it as a standalone Node.js bundle you run entirely inside your own AWS, GCP, or on-prem environment. Evicta's cloud never sees a single byte — not the credentials, not the tickets, not the output. This is included in the $1,499 Premium tier. The only scenario where DIY is the only path is if your policy prohibits running any externally-generated code whatsoever.
Frequently Asked Questions
What happens to my data when I cancel Zendesk?
The moment you cancel, you lose login access immediately. Your data enters a 90-day grace window during which the account owner can reactivate. After 90 days, Zendesk permanently and irreversibly deletes all structured service data — tickets, comments, custom fields, users, and organizations. There is no recovery. Export everything before you cancel.
Can I export Zendesk data after canceling?
No. Zendesk's own documentation is explicit: you cannot sign in or access account data after cancellation. If you cancel before exporting, you have at most a 90-day window to reactivate and retrieve your data — but reactivation requires billing, and Zendesk does not guarantee it will be honored. Don't rely on the grace window.
What does Zendesk's native CSV export miss?
Quite a lot. Native CSV export excludes multi-line text fields, multi-select custom fields, custom date fields, ticket comments, custom user fields, and custom organization fields. Sunshine Custom Objects are not included at all. For a complete export you need to use the Incremental Export API and the Custom Objects API directly — or a tool like Evicta that wraps both.
What is the July 1, 2026 Sunshine deadline?
Zendesk is permanently shutting down its Sunshine Custom Objects API on July 1, 2026. After that date, the API returns 410 Gone and the data is gone. This is independent of your subscription status — even if you keep paying, the Sunshine API stops working. If your Zendesk instance uses Sunshine for hardware assets, license entitlements, B2B account hierarchies, or any relational data model, you must extract those records before July 1, 2026.
How long does it take to extract Zendesk data with Evicta?
Under 24 hours for any volume up to 2 million records. The extraction runs against Zendesk's Incremental Export API with automatic rate-limit handling, cursor state tracking, and schema validation. You review a 100-record preview before the full run starts. No card required for the preview.
Does Evicta pick the new help desk for me?
No. Evicta extracts your Zendesk data and delivers it as Postgres or JSONL. Where you take that data — Freshdesk, Intercom, HubSpot, Salesforce, your own data warehouse — is your call. We're the extraction layer, not the destination.