Skip to content
DataGrine home

Helpdesk migration checklist: before, during and after the move

A helpdesk migration goes well when it happens in order: take stock of what you have, decide in writing where every field lands, prepare the destination, check a sample of your own records there, run the full migration while your team keeps working, and cut over with a final catch-up run. This checklist covers each stage.

By the DataGrine team · Updated · 3 min read

Key takeaways

  • Configuration such as triggers, macros, views and SLAs is not data. List it before you start, because it is rebuilt by hand.
  • Every field and dropdown value needs a written destination, or a deliberate decision to leave it behind.
  • Test a sample of your hardest records: long threads, attachments, inline images, many custom fields.
  • Count records on both sides afterwards, and explain every difference.

1. Take stock of what you have

  • Record counts for every object: companies, contacts, tickets or conversations, knowledge base articles.
  • The date range you actually need.
  • Every custom field, and which ones anyone still fills in or reports on.
  • Agents and groups, including people who have left but still own old tickets.
  • Knowledge base structure and the languages articles are translated into.
  • Integrations that use the old helpdesk’s API. They share its rate limit with the migration.
  • Voice history, if you run a contact centre. Migrating call logs and recordings needs its own plan: recording links expire, and retention clocks keep running.
  • Configuration: triggers, automations, macros, views, SLAs, business hours. This is the list you will rebuild.

2. Decide the scope

Decide which objects move, which records move (all history or a date range, with spam and deleted tickets left out), and what happens to what stays behind: an export kept for compliance, or the old helpdesk kept read-only for a while.

3. Map every field and value in writing

  • Statuses to states, one by one, including "pending" and "on hold".
  • Agents to teammates and groups to teams, with a default for anyone who does not exist on the destination.
  • A fallback for every value with no counterpart: a custom field, a tag, a note, or a deliberate drop.
  • How contacts are matched: on email, external ID or phone, never on name.
  • Dates: check whether the destination accepts an original creation date on new tickets. Where it does not, carry it in a field of its own.

4. Prepare the destination

  • Create teams, agents and groups, so assignment has somewhere to land.
  • Create the custom fields, ticket types and states the mapping points at.
  • Switch on the knowledge base languages you will carry.
  • Check what the destination sends customers during an import, and pause automations that react to new tickets. Zendesk, for example, documents that triggers do not run on imported tickets unless they are closed.
  • Decide where old help centre URLs will redirect.

5. Run a sample and sign it off

Move a small set of your own records, chosen because they are hard, and check them in the destination:

  • Threads arrive whole and in order, with the right author on each reply.
  • Internal notes stay internal.
  • Attachments arrive, and inline images still show.
  • Assignment, status, priority and custom fields match the mapping, and resolved tickets arrive resolved.
  • One named person on your side signs the sample off before the full run.

6. Run the full migration, then cut over

  1. 1.Run the full migration while the team keeps working in the old helpdesk. Pause heavy integrations if you can.
  2. 2.Run a delta: only records created or changed since the full run.
  3. 3.Switch the channels: email forwarding, chat widget, phone and social.
  4. 4.Turn on the help centre redirects and tell the team which helpdesk to answer in, from what time.

7. Reconcile, then rebuild what is left

  • Count records on both sides, object by object, and explain every difference.
  • Spot-check your biggest customers and longest threads.
  • Rebuild the configuration you listed in step 1.
  • Keep the old helpdesk readable until the reconciliation is signed off.

Migrating with DataGrine

These stages are built in: read-only access to the source, a written field mapping, a sample run signed off before the full run, a delta sync at cut-over, rollback per run and an itemised reconciliation report. Run it self-serve on the DataGrine platform or have our migration team handle it end to end, with 24/7 support either way. See how a migration runs.

Questions

Asked about this, often.

Do we need to stop support during a helpdesk migration?

No. The full migration runs while your team keeps working in the old helpdesk, and a delta run at cut-over moves only what was created or changed since. The switch itself is short.

What usually goes wrong in a helpdesk migration?

Internal notes arriving as public replies, inline images that stop showing, resolved tickets arriving open, and original dates being lost. Each is caught by checking a sample run before the full migration.

Who should sign off a migration?

One named person on the support side signs off the field mapping and the sample run before the full migration, and the reconciliation after it.

Let’s move your data forward.

Tell us where you are moving from and to. Run it yourself on the DataGrine platform or let our migration team handle it end to end, with 24/7 support either way.

  • A written mapping document before anything moves
  • A sample migration you review before the live run
  • Delta sync at cut-over and rollback per run
  • 24/7 premium support until you sign off