Knowledge — data migration
Migrating from Excel to housing community software: reconcile first, publish second
The main migration risk is not XLSX or CSV format. It is carrying ambiguity into a new system and making that ambiguity look orderly.
1. Inventory all sources
List every source containing relevant data: unit sheets, owner lists, ownership shares, balances, funds, meters, documents and open cases.
For every source record who owns it, when it was last current and what it covers. Where two files describe the same state, choose the canonical source.
2. Set a cut-off point
Without a cut-off, the old and new systems can change the same state at the same time. Set the date and time after which updates are recorded in one place, or define a controlled synchronisation procedure.
3. Map meaning, not only column names
Define the meaning of:
- unit identifier;
- owner and representation;
- ownership share;
- effective date;
- balance at cut-off;
- fund type;
- meter identifier;
- case status.
A “balance” column without an “as of” date is not sufficient.
4. Resolve ambiguity before publication
Stop records where there is, for example:
- the same unit under two identifiers;
- an owner without an unambiguous unit link;
- conflicting ownership shares;
- missing effective date;
- balance totals that do not reconcile to the control source;
- uncertainty whether a field represents zero or missing data;
- mixed number/date formats that can be interpreted differently.
The system should not guess which version is correct.
5. Run a dry import
The dry import should produce preview and exception reports without changing the operative state.
Check:
- number of units;
- number of owners;
- ownership-share total where applicable;
- balances at the agreed cut-off;
- rejected records;
- duplicates;
- records missing required relationships.
6. Pilot one community
When a firm moves several communities, the first one should be a pilot. Lessons from mapping and exception handling can then improve subsequent migrations.
Do not choose only the easiest data set. A useful pilot should contain several representative data problems.
7. Reconcile the first operating cycle
Migration does not end with “import successful”. After the first charging or settlement cycle, compare the output with an independent reference and investigate differences.
STOP conditions
- unit counts cannot be reconciled;
- unresolved owner/share conflicts remain;
- opening balances do not have a confirmed date;
- control totals differ without explanation;
- field meaning has to be guessed;
- source files and mapping have not been preserved.
Definition of Done
Migration is complete when:
- record counts reconcile;
- control totals match or differences are explicitly accepted;
- every exception has an owner and status;
- source files and mapping are preserved;
- the approver and publication time are known;
- the first operating cycle passed reconciliation;
- the new system can export the migrated data again.
Sources and scope
- WM Administrator privacy policy
- WM Administrator technical and organisational measures
- WM Administrator subprocessors and data recipients
For information only. This is not legal, tax or accounting advice; adapt the procedure to the documents and circumstances of the specific housing community.