Key points
- Preserve stable identifiers and document every source-to-target mapping.
- Separate active records, useful history and exclusions under an approved policy.
- Reconcile open tasks and approve discrepancies before cutover.
1. Decide what genuinely needs to move
The executive housekeeper and IT lead list the operational objects required: rooms, buildings, zones, users, roles, relevant equipment references and tasks that remain active. Each dataset has a business owner. The supplier confirms accepted formats and importable objects; a field in the old application does not guarantee an equivalent in the new one.
For historical records, choose transfer, approved archiving or exclusion. An old comment may contain personal information with no current purpose. Archive decisions should specify access, storage and the retention period set by the applicable policy. Data cleaning does not authorise arbitrary deletion of documents the hotel needs to retain.
2. Build the source-to-target mapping table
| Object | Source | Destination and transformation | Check |
|---|---|---|---|
| Room | Internal identifier and displayed number | Destination identifier; preserve meaningful leading zeros | One mapping per room |
| Zone | Floor or building label | Approved naming convention | No active room without a zone |
| User | Account and current employment status | Authorised account and reviewed role | Manager approval |
| Referenced equipment | Code and location | Confirmed importable reference | No incorrect room association |
| Open task | Identifier, state and owner | Destination state and owner | Explicit continuation of work |
Add a rule for missing values. An empty cell must not silently turn into “completed” or “no issue”. Document the time zone and date format when they determine a deadline or explain the sequence of events. These details matter when the source and destination display timestamps differently.
3. Correct records with a decision trail
Work on an authorised copy and protect the export. For each correction, record the previous content, replacement value, reason and approver. Resolve duplicates using identifiers and context, rather than similar names alone. Two rooms numbered “12” in different buildings are not necessarily duplicates, and merging them could attach work to the wrong location.
Fictional example: a property has 80 active rooms. Its export contains 82 rows: the 80 rooms, a former room now used for storage and an accidental duplicate. The manager confirms the 80 rooms to transfer; the storage location is handled separately according to available destination objects; the duplicate is excluded with a recorded explanation. Importing all 82 and hiding the errors would preserve uncertainty in future assignments.
4. Run a trial import and reconcile the outcome
The supplier performs a trial in the agreed environment. Compare totals and relationships: a task attached to the wrong room can pass a simple record-count check. Inspect exceptional cases, building changes, accented characters and tasks with attachments when attachment transfer is included. Keep a discrepancy list that distinguishes rejected rows from records accepted with altered content.
CNIL recommends checking backup integrity. Keep a usable reference copy for reconciliation and agree restoration options with the suppliers. A successful export does not, by itself, prove that the previous service can be returned to operation. Confirm that separate question before making the final switch.
5. Close discrepancies before daily use
At the planned cutover, identify changes made since the trial and transfer that difference using the agreed procedure. Every open task receives a destination identifier or an explicit handling instruction outside the import. Housekeeping checks allocation by zone, maintenance confirms its outstanding jobs, and the administrator checks permissions.
Close the migration when discrepancies have been resolved or explicitly accepted, active tasks have owners and the final reference has been archived through an authorised process. This data preparation supports digitising housekeeping operations, but training staff and validating interface behaviour remain separate responsibilities.
Frequently asked questions
Should every historical comment be transferred?
No. Assess operational usefulness, retention requirements and any personal information. Approve the transfer rules before producing the final export.
What if a source value has no destination field?
Record the gap and agree an authorised approach with its business owner: a confirmed compatible field, an accessible archive or separate handling. Do not repurpose a field in a way that makes its meaning misleading.
