Migrating a service desk is mostly a data and workflow exercise, not a software installation. The sequence that keeps risk low is the same every time: map the queues and workflows in use, validate the data, migrate in a controlled window, and verify against the original before switching the intake channel.
For many teams, the main obstacle to leaving a community ticketing tool is not the cost of the new platform. It is the fear of losing what the old one holds: years of ticket history, queues that took time to tune and service levels people rely on. This guide answers that objection. It explains what is actually being migrated, lays out the four-phase sequence with the deliverable and exit criterion of each phase, lists what usually breaks and how to catch it early, and gives a realistic view of how long a migration takes.
What you are actually migrating
It helps to stop thinking of a migration as moving “the software” and start thinking of it as moving a working service. A ticketing tool holds data, but it also holds the rules that make the service behave the way users and technicians expect. Six elements make up the scope:
- Tickets and their history. Every ticket with its articles, notes, status changes and attachments. History is usually the first requirement people ask about.
- Customers and users. The people who open tickets, the technicians who handle them, and the roles and permissions that separate the two.
- Queues and ownership. The structure that routes work to the right team, and who is responsible for each queue.
- Service levels and calendars. Response and resolution targets, working hours and time zones, which are easy to get subtly wrong.
- Workflows and automations. Notifications, escalations, auto-assignments and templates that run without anyone touching them.
- Integrations. Connections with email, monitoring, directories and other systems that create or update tickets.
The table below turns each element into a question to answer before anything moves, and shows what happens when the question is skipped:
| Element | Question to answer first | Risk if it is skipped |
|---|---|---|
| Tickets and history | How much history is needed, and in what format does it need to be searchable? | Teams lose the ability to look up past cases. |
| Users and permissions | Which roles exist today, and which should exist tomorrow? | People see too much or too little after the switch. |
| Queues | Which queues are active and who owns each one? | Tickets land in queues nobody watches. |
| Service levels | Which targets and calendars apply to which queues? | Reports show breaches that never happened, or hide real ones. |
| Workflows | Which automations are actually used? | Dead rules are carried over, and live ones are missed. |
| Integrations | Which systems create or update tickets? | Email or monitoring stops reaching the service desk. |
The four-phase sequence
The sequence that keeps risk low is the same every time. Each phase ends with something concrete and a criterion that says whether it is safe to continue:
| Phase | What happens | Deliverable | Exit criterion |
|---|---|---|---|
| 1. Map | Inventory the queues, workflows, customisations and integrations in use. | A scope document listing what moves, what is rebuilt and what is retired. | The business owner of each queue has confirmed it. |
| 2. Validate | Check the data for quality, duplicates and gaps, and run a test migration. | A validated data set and a test-migration report. | The test results match the original within the agreed tolerance. |
| 3. Migrate | Move the data in a controlled window, with open tickets carried with their state. | The production data in the new platform. | The migration completes with no unresolved errors. |
| 4. Verify | Compare the new platform against the original before switching the intake channel. | A verification checklist, signed off. | Counts, queues, permissions, SLAs and integrations all match; only then does intake switch. |
Phase 1: map. The most valuable output of the first phase is a decision about what not to move. Every queue, automation and integration is classified as migrate, rebuild or retire. Teams that skip this step carry years of accumulated clutter into the new system.
Phase 2: validate. Data from a community tool is rarely perfectly clean. Duplicate customers, orphaned attachments, inconsistent statuses and custom fields that mean different things to different teams are common. A test migration surfaces these issues while there is still time to fix them. OTRS describes its own approach in the same terms: validate workflows and data first, then migrate, and it recommends scheduling a test migration beforehand.
Phase 3: migrate. The migration itself should happen in a controlled window, with a clear plan for tickets that are open at the moment of the move. They are migrated with their state, so a ticket that was waiting for a customer reply is still waiting for a customer reply. OTRS states that it transfers all relevant data, including users, tickets, attachments and history, and that its experts confirm this in advance.
Phase 4: verify. Before the intake channel is switched, the new platform is compared against the original: ticket counts per queue, permissions, service-level calculations and each integration. Only after that checklist is signed off does the email address, portal or form that users rely on point to the new system. Keeping the old system available in read-only mode for a period gives the team a way back if something was missed.
What usually breaks, and how to catch it early
Migrations rarely fail because of one dramatic problem. They fail because of several small mismatches that nobody tested. Seven are worth checking on purpose:
| What breaks | Why it happens | How to catch it early |
|---|---|---|
| Custom fields | Fields in the old tool do not map one-to-one to the new structure. | Map every custom field in phase 1 and check a sample of tickets in the test migration. |
| Attachments | Large or orphaned files are skipped or fail silently. | Compare attachment counts and sizes before and after. |
| Ticket numbers and links | Old ticket numbers quoted in emails or other systems stop resolving. | Decide how old numbers are kept or cross-referenced before the move. |
| Service-level calendars | Working hours and time zones are configured differently. | Recalculate a sample of tickets and compare the due dates. |
| Duplicate notifications | Automations fire in both systems, or twice in the new one. | Disable old notifications at cutover and test with a small group. |
| Integrations | Credentials, endpoints or formats change. | Test each integration end to end in the test migration. |
| Permissions | Roles are mapped too broadly or too narrowly. | Review a sample of users from each role in the new system. |
On integrations in particular, OTRS notes that interfaces to third-party systems can in most cases be carried over once they have been checked and, if necessary, adapted. The word “checked” is doing real work in that sentence: an integration that is assumed to work is the most common cause of a surprise on the first day.
The most common mistake is not technical. It is migrating workflows nobody uses any more. A migration is a rare opportunity to clean up, and the teams that treat it as a copy operation end up paying to maintain rules that stopped being relevant years ago.
How long it realistically takes
There is no single answer, and anyone who offers one without looking at the environment is guessing. Duration depends on queue count, integrations and data quality rather than on the size of the organisation. A company with thousands of employees and three clean queues can migrate faster than a small one with a dozen queues, years of custom fields and five integrations.
Four factors explain most of the variation:
- Number of queues and workflows. Each one needs a decision and a test.
- Number and complexity of integrations. Each one needs to be checked end to end.
- Data quality. Clean data moves quickly; messy data needs remediation first.
- Tolerance for downtime. A migration that can afford a maintenance window is simpler than one that must run without interruption.
For reference, OTRS publishes effort figures for its own migration options, which depend on the size and customisation of each installation. An on-premises migration is described as taking two to three days, with a test migration recommended and downtime planned. A migration to its managed service is described as one to two days including a test migration, with downtime required. Setting up a new instance with ticket transfer is described as four to six consulting days without downtime, and a predefined solution scenario with standard configuration as two consulting days without downtime. These are OTRS’s figures for its own offerings and are a reference point, not a promise for any other project.
The lower-risk path is almost always incremental. Starting with one queue, verifying it fully and then expanding gives the team early evidence and limits the impact of any surprise. It also answers the question of whether everything has to move at once: it does not.
A practical way to begin
A 20-minute diagnostic conversation is enough to place an environment on this map. It covers which tool is running today, how many queues and integrations depend on it, how much history matters and what level of downtime is acceptable. The aim is to leave with a clear view of the sequence, the likely effort and the first queue to move, not with a sales commitment.
Aufiero Informática is the official OTRS distributor in Latin America, with local contracts and support in Portuguese and Spanish. The current OTRS platform runs as SaaS, managed or self-hosted, so the choice of where it runs stays with the organisation.
Frequently asked questions
Can ticket history be migrated?
Yes, when the data is validated before the move. History is usually the first requirement to confirm.
How long does a migration take?
It depends on queue count, integrations and data quality rather than on the size of the organisation.
Do we have to migrate everything at once?
No. Starting with one queue and expanding is the lower-risk path.
What happens to open tickets during the switch?
They are migrated with their state; the intake channel switches only after verification.
What is the most common mistake?
Migrating workflows nobody uses any more, instead of taking the move as a chance to clean them up.
Aufiero Informática is the official OTRS distributor in Latin America.

